CN116151862A

CN116151862A - Data processing method, related device, equipment and storage medium

Info

Publication number: CN116151862A
Application number: CN202111363510.1A
Authority: CN
Inventors: 王继天; 周梦; 魏晔纯; 孟程程
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2021-11-17
Filing date: 2021-11-17
Publication date: 2023-05-23

Abstract

The application discloses a data processing method, a related device, equipment and a storage medium, wherein the method comprises the following steps: acquiring first object data of an object to be predicted in a first time period; characterizing the first object data to determine first data characteristics of the object to be predicted; acquiring a first predicted data tolerance of an object to be predicted through a data tolerance prediction model based on first data characteristics of first object data; and acquiring the first predicted object tolerance of the object to be predicted in the second time period through the object tolerance prediction model based on the first predicted data tolerance and the first object tolerance of the object to be predicted. According to the method, the predicted object tolerance of the object in the next time period can be obtained based on the predicted data tolerance and the object tolerance, and the problem of low accuracy caused by the questionnaire is avoided on the basis of considering the object tolerance of the object, so that the accuracy of predicting the object tolerance of the object is improved.

Description

Data processing method, related device, equipment and storage medium

Technical Field

The present disclosure relates to the field of data processing technologies, and in particular, to a data processing method, a related device, an apparatus, and a storage medium.

Background

In recent years, with the continuous expansion of the business scale of financial institutions such as commercial banks and the continuous expansion of the financial business scope, the financial institutions can provide different investment products for the objects, and before the objects purchase the investment products, the banks or the financial institutions need to evaluate the tolerance of the investment objects, divide the tolerance of the investment objects of the objects into different investment types according to the evaluation results, and the objects can purchase the investment products based on the investment types corresponding to the tolerance of the investment objects. However, when a subject fills out a questionnaire, there may be a behavior deviation, and subjective judgment is made on the content of the questionnaire by the subject at the time of filling out the questionnaire, thus reducing the accuracy of the evaluation result obtained by the questionnaire. Therefore, how to improve the accuracy of the object tolerance prediction result is a problem to be solved.

Disclosure of Invention

The embodiment of the application provides a data processing method, a related device, equipment and a storage medium, wherein the data characteristics of an object to be predicted are constructed through information of multiple dimensions related to the object to be predicted in a time period, the predicted data tolerance is obtained through a data tolerance prediction model based on the constructed data characteristics, and the accuracy of the predicted data tolerance is ensured. And further, based on the predicted data tolerance and the object tolerance with higher accuracy, the predicted object tolerance of the object to be predicted in the next time period is obtained through the object tolerance prediction model, and the problem of low accuracy caused by the behavior deviation of filling the questionnaire by the object is avoided on the basis of considering the object tolerance of the object to be predicted, so that the accuracy of object tolerance prediction on the object is improved.

In view of this, a first aspect of the present application provides a method of data processing, comprising:

acquiring first object data of an object to be predicted in a first time period, wherein the first object data comprises first object tolerance and first object information, the first object tolerance is used for representing the tolerance degree of the object to be predicted, which is reduced in the amount of resources corresponding to products owned by the object to be predicted in the first time period, and the first object information is related behavior information of the object to be predicted on the products in the first time period;

characterizing the first object data to determine first data characteristics of the object to be predicted;

acquiring a first predicted data tolerance of an object to be predicted through a data tolerance prediction model based on first data characteristics of first object data;

and acquiring the first predicted object tolerance of the object to be predicted in a second time period based on the first predicted data tolerance and the first object tolerance of the object to be predicted, wherein the second time period is the next time period adjacent to the first time period, and the first predicted object tolerance is used for representing the prediction tolerance degree of the object to be predicted, which is reduced in the amount of resources corresponding to the product owned by the object to be predicted in the second time period.

A second aspect of the present application provides a data processing apparatus comprising:

the system comprises an acquisition module, a prediction module and a prediction module, wherein the acquisition module is used for acquiring first object data of an object to be predicted in a first time period, the first object data comprises first object tolerance and first object information, the first object tolerance is used for representing the tolerance degree of the object to be predicted, which is the tolerance degree of the object to be predicted, of the reduction of the corresponding resource amount of a product owned by the object to be predicted in the first time period, and the first object information is related behavior information of the object to be predicted, which is generated to the product in the first time period;

the determining module is used for carrying out characterization processing on the first object data and determining first data characteristics of the object to be predicted;

the acquisition module is further used for acquiring a first predicted data tolerance of the object to be predicted through the data tolerance prediction model based on the first data characteristic of the first object data;

the obtaining module is further configured to obtain, according to the first prediction data tolerance of the object to be predicted and the first object tolerance, a first prediction object tolerance of the object to be predicted in a second time period through the object tolerance prediction model, where the second time period is a next time period adjacent to the first time period, and the first prediction object tolerance is used to characterize a prediction tolerance degree of a decrease in a resource amount corresponding to a product owned by the object to be predicted in the second time period.

In one possible implementation, the obtaining module is further configured to obtain market information in a first time period;

the determining module is further used for determining the market prediction influence degree of the object to be predicted based on the market information and the first object data in the first time period, wherein the market prediction influence degree indicates the influence of the trading market on the behavior of the object to be predicted in the first time period;

the obtaining module is specifically configured to obtain, by using an object tolerance prediction model, a first predicted object tolerance of the object to be predicted in a second time period based on a first predicted data tolerance of the object to be predicted, the first object tolerance, and a market prediction influence of the object to be predicted.

In one possible embodiment, the first object data further comprises object base information;

the determining module is specifically used for determining first data characteristics of the object to be predicted and object image characteristics of the object to be predicted based on the characteristic processing of the first object data;

the acquisition module is specifically configured to acquire a first predicted data tolerance of an object to be predicted through the data tolerance prediction model based on a first data feature of the first object data and an object portrait feature of the object to be predicted.

In one possible implementation manner, the obtaining module is further configured to obtain second object data of the object to be predicted in a second time period, where the second object data includes a second object tolerance and second object information, the second object tolerance is used to characterize a tolerance degree of the object to be predicted for a decrease in a resource amount corresponding to a product owned by the object to be predicted in the second time period, and the second object information is related behavior information generated by the object to be predicted for the product in the second time period;

the determining module is also used for carrying out characterization processing on the second object data and determining the second data characteristics of the object to be predicted;

the acquisition module is further used for acquiring second prediction data tolerance of the object to be predicted through the data tolerance prediction model based on second data characteristics of the second object data;

the obtaining module is further configured to obtain, based on a second predicted data tolerance of the second object data and the second object tolerance, a second predicted object tolerance of the object to be predicted in a third time period through the object tolerance prediction model, where the third time period is a next time period adjacent to the second time period, and the second predicted object tolerance is used to characterize a predicted tolerance degree of a decrease in a resource amount corresponding to a product owned by the object to be predicted in the third time period.

In one possible implementation, the first predicted object tolerance is the second object tolerance.

In one possible embodiment, the first object tolerance is determined based on a questionnaire filled in by the object to be predicted in the first time period;

the determining module is further configured to determine a third object tolerance based on the questionnaire updated by the object to be predicted in the second time period if the object to be predicted updates the questionnaire in the second time period;

the determining module is further configured to determine the third object tolerance as the second object tolerance if the third object tolerance is less than the first predicted object tolerance.

In one possible embodiment, the second object information includes second object transaction information, or the second object information includes second object transaction information and second object access information;

the determining module is further used for determining the first data tolerance of the object to be predicted in the second period through a redemption strategy based on second object transaction information in the second object information;

the determining module is further configured to determine the first data tolerance as a second object tolerance.

In one possible implementation, the second object information includes second object holding information, or the second object information includes second object holding information and second object access information;

The determining module is further used for determining second data tolerance of the object to be predicted in a second period based on second object holding information in the second object information;

the determining module is further configured to determine the second data tolerance as a second object tolerance if the second data tolerance is less than the first predicted object tolerance.

In one possible embodiment, the data features include at least one of an access feature, a purchase feature, and a holding feature, the access feature being based on the object access information, the purchase feature being based on the object transaction information, and the holding feature being based on the object holding information;

the data tolerance prediction model comprises an access data tolerance prediction model, a buying data tolerance prediction model and a holding data tolerance prediction model;

the access data tolerance prediction model is used for obtaining access prediction data tolerance of the object based on the access characteristics;

the purchasing data tolerance prediction model is used for obtaining the purchasing prediction data tolerance of the object based on the purchasing characteristics;

the holding data tolerance prediction model is used for obtaining the holding prediction data tolerance of the object based on the holding characteristics;

the first prediction data tolerance of the object to be predicted comprises at least one of access prediction data tolerance of the object to be predicted, purchase prediction data tolerance of the object to be predicted and holding prediction data tolerance of the object to be predicted.

In one possible implementation, the data processing apparatus further includes a screening module and a training module;

the system comprises an acquisition module, a storage module and a storage module, wherein the acquisition module is further used for acquiring an initial data sample set, wherein the initial data sample set comprises initial data samples of a plurality of object samples, and each initial data sample comprises at least one of object tolerance, object basic information, object access information, object bin holding information and object transaction information;

a screening module for screening a set of target data samples from the initial set of data samples based on the redemption policy and the holding policy, wherein the set of target data samples includes target data samples of the plurality of object samples;

a determination module for determining a target data tolerance for each of the object samples based on the set of target data samples by redemption policies and the holding policies;

the acquisition module is also used for acquiring the predicted data tolerance of each object sample through the data tolerance prediction model to be trained based on the target data sample set;

the training module is used for training the data tolerance prediction model to be trained based on the target data tolerance of each object sample and the predicted data tolerance of each object sample so as to obtain the data tolerance prediction model.

In one possible implementation manner, the obtaining module is further configured to obtain, based on the predicted data tolerance of each object sample and the object tolerance of each object sample, a predicted object tolerance of each object sample through an object tolerance prediction model to be trained;

the training module is further configured to train the object tolerance prediction model to be trained based on the object tolerance of each object sample and the predicted object tolerance of each object sample, so as to obtain the object tolerance prediction model.

In one possible implementation, the obtaining module is further configured to obtain a market information set, where the market information set includes market information in a plurality of different periods;

the determining module is further used for determining the market prediction influence degree of each object sample in different periods based on the market information in each different period and the target data samples of the plurality of object samples, wherein the market prediction influence degree indicates the influence of the market on the behaviors of the object samples in one time period;

the obtaining module is specifically configured to obtain, according to the object tolerance prediction model to be trained, a predicted object tolerance of each object sample based on the predicted data tolerance of each object sample, the object tolerance of each object sample, and the market prediction influence of each object sample in different periods.

A third aspect of the present application provides a computer readable storage medium having instructions stored therein which, when run on a computer, cause the computer to perform the method of the above aspects.

A fourth aspect of the present application provides a computer device, comprising: memory, transceiver, processor, and bus system; wherein the memory is used for storing programs; the processor is used for executing the program in the memory to realize the method in the aspects; the bus system is used for connecting the memory and the processor so as to enable the memory and the processor to communicate.

A fifth aspect of the present application provides a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the method described in the above aspects.

From the above technical solutions, the embodiments of the present application have the following advantages:

In the embodiment of the present application, a method for processing data is provided, first object data of an object to be predicted in a first time period is obtained, where the first object data includes a first object tolerance and first object information, the first object tolerance is used to characterize a tolerance degree of a decrease in a resource amount corresponding to a product owned by the object to be predicted in the first time period, and the first object information is related behavior information generated by the object to be predicted on the product in the first time period. And finally, based on the first predicted data tolerance of the object to be predicted and the first object tolerance, acquiring the first predicted object tolerance of the object to be predicted in a second time period, wherein the second time period is the next time period adjacent to the first time period, and the first predicted object tolerance is used for representing the prediction tolerance degree of the object to be predicted, which is reduced in the amount of resources corresponding to products possessed by the object to be predicted in the second time period. According to the method, the data characteristics of the object to be predicted are constructed through the information of the plurality of dimensions related to the object to be predicted in a time period, the predicted data tolerance is obtained through the data tolerance prediction model based on the constructed data characteristics, and the accuracy of the predicted data tolerance is ensured. And further, based on the predicted data tolerance and the object tolerance with higher accuracy, the predicted object tolerance of the object to be predicted in the next time period is obtained through the object tolerance prediction model, and the problem of low accuracy caused by the behavior deviation of filling the questionnaire by the object is avoided on the basis of considering the object tolerance of the object to be predicted, so that the accuracy of object tolerance prediction on the object is improved.

Drawings

FIG. 1 is a schematic diagram of a system for data processing according to an embodiment of the present disclosure;

FIG. 2 is a flow chart of a method for data processing according to an embodiment of the present disclosure;

FIG. 3 is a schematic diagram of one embodiment of a method for data processing according to an embodiment of the present application;

FIG. 4 is a schematic flow chart of predicting tolerance of an object based on market information according to an embodiment of the present disclosure;

FIG. 5 is a schematic flow chart of obtaining a predicted data tolerance based on object representation features according to an embodiment of the present application;

FIG. 6 is a flow chart of predictive data tolerance determination provided in an embodiment of the present application;

fig. 7 is a schematic flow chart of training a data tolerance prediction model to be trained according to an embodiment of the present application;

fig. 8 is a schematic flow chart of training a data tolerance prediction model to be trained according to an embodiment of the present application;

FIG. 9 is a schematic diagram of a data processing apparatus according to an embodiment of the present application;

FIG. 10 is a schematic diagram of an embodiment of a server according to an embodiment of the present application;

fig. 11 is a schematic diagram of an embodiment of a terminal device in an embodiment of the present application.

Detailed Description

The terms "first," "second," "third," "fourth" and the like in the description and in the claims of this application and in the above-described figures, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that embodiments of the present application described herein may be capable of operation in sequences other than those illustrated or described herein, for example. Furthermore, the terms "comprises," "comprising," and "includes" and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.

In recent years, with the continuous expansion of the business scale of financial institutions such as commercial banks and the continuous expansion of the financial business scope, the financial institutions can provide different investment products for the objects, and before the objects purchase the investment products, the banks or the financial institutions need to evaluate the tolerance of the investment objects, divide the tolerance of the investment objects of the objects into different investment types according to the evaluation results, and the objects can purchase the investment products based on the investment types corresponding to the tolerance of the investment objects. However, when a subject fills out a questionnaire, there may be a behavior deviation, and subjective judgment is made on the content of the questionnaire by the subject at the time of filling out the questionnaire, thus reducing the accuracy of the evaluation result obtained by the questionnaire. Based on the above, the embodiment of the application provides a data processing method, which obtains the predicted object tolerance of the object to be predicted in the next time period through the object tolerance prediction model, and avoids the problem of low accuracy caused by the behavior deviation of filling the questionnaire by the object on the basis of considering the object tolerance of the object to be predicted, thereby improving the accuracy of object tolerance prediction on the object.

First, for ease of understanding, some terms or concepts related to the embodiments of the present application are explained first.

1. Tolerance of

In this embodiment, the object tolerance and the data tolerance are both the cost loss degrees that the object can accept when buying the investment product on the financial platform. For example, the redemption of the object at principal 1000 and principal loss 100 may indicate that the object tolerance or data tolerance of the object is 10% (100/1000 x 100% = 10%).

2. Multiple linear regression model

Since the variation of the socioeconomic phenomenon is often affected by a plurality of factors, multiple regression analysis is performed, and regression including two or more independent variables is called multiple linear regression, and the regression model thus constructed becomes a multiple linear regression model.

The application scenario of the embodiment of the present application is described below. It will be appreciated that the method of data processing may be performed by the terminal device or by the server. Referring to fig. 1, fig. 1 is a schematic diagram of a system of a data processing method provided in an embodiment of the present application, as shown in fig. 1, where the video processing system includes a server and a terminal device, and when the data processing method is deployed on the terminal device, the terminal device can obtain a predicted object tolerance of an object to be predicted in a next time period through the data tolerance prediction model and the object tolerance prediction model by using a plurality of dimensions of information related to the object to be predicted in a time period in an offline state after training of the data tolerance prediction model and the object tolerance prediction model is completed by the terminal device, and the terminal device does not need to be connected with the network at this time, so that the object tolerance prediction process is more convenient. When the data processing method is deployed on the server, the server can be based on hardware performance of the server and information of multiple dimensions related to the object to be predicted, which is acquired by different channels in cloud storage, so that the information of the object to be predicted is richer and more detailed, and meanwhile, the object tolerance prediction can be improved, and therefore, the object tolerance obtaining efficiency can be improved under the condition of ensuring the accuracy of the object tolerance prediction.

It should be noted that, the server in fig. 1 may be an independent physical server, or may be a server cluster or a distributed system formed by a plurality of physical servers, or may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms. The terminal device may be a tablet computer, a notebook computer, a palm computer, a mobile phone, a personal computer (personal computer, PC) and an intelligent voice interaction device shown in fig. 1, and the terminal device may also include, but is not limited to, an intelligent home appliance, a vehicle-mounted terminal and the like. And the terminal device and the server may be directly or indirectly connected through a wireless network, a wired network, or a removable storage medium. Wherein the wireless network uses standard communication techniques and/or protocols. The wireless network is typically the internet, but may be any network including, but not limited to, bluetooth, a local area network (Local Area Network, LAN), a metropolitan area network (Metropolitan Area Network, MAN), a wide area network (Wide Area Network, WAN), a mobile, private network, or any combination of virtual private networks. In some embodiments, custom or dedicated data communication techniques may be used in place of or in addition to the data communication techniques described above. The removable storage medium may be a universal serial bus (Universal Serial Bus, USB) flash drive, a removable hard disk, or other removable storage medium, etc.

Second, although only five terminal devices and one server are shown in fig. 1, it should be understood that the example in fig. 1 is only for understanding the present solution, and the number of specific terminal devices and servers should be flexibly determined according to actual situations.

The training process of the data tolerance prediction model and the object tolerance prediction model in the embodiment of the present application needs to be implemented based on the artificial intelligence field, and some basic concepts of the artificial intelligence field are described before the description of the data processing method provided in the embodiment of the present application begins. Artificial intelligence (Artificial Intelligence, AI) is the theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use the knowledge to obtain optimal results. In other words, artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence, i.e. research on design principles and implementation methods of various intelligent machines, enables the machines to have functions of sensing, reasoning and decision. The artificial intelligence technology is a comprehensive subject, and relates to the technology with wide fields, namely the technology with a hardware level and the technology with a software level. Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and other directions.

With research and progress of artificial intelligence technology, the artificial intelligence technology is developed in various directions, and Machine Learning (ML) is a multi-domain interdisciplinary subject, and relates to multiple subjects such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. It is specially studied how a computer simulates or implements learning behavior of a human to acquire new knowledge or skills, and reorganizes existing knowledge structures to continuously improve own performance. Machine learning is the core of artificial intelligence, a fundamental approach to letting computers have intelligence, which is applied throughout various areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, confidence networks, reinforcement learning, transfer learning, induction learning, teaching learning, and the like.

For ease of understanding, referring to fig. 2, fig. 2 is a schematic flow chart of a method for data processing according to an embodiment of the present application, and referring to fig. 2, the flow chart of the method for data processing is divided into three stages, specifically, a data tolerance prediction model training stage, an object tolerance prediction model training stage, and an object tolerance prediction stage. The following describes the functions and processes of each stage, specifically:

In the data tolerance prediction model training stage A1, an initial data sample set is firstly obtained, wherein the initial data sample set comprises initial data samples of a plurality of object samples, and each initial data sample comprises at least one of object tolerance, object basic information, object access information, object bin holding information and object transaction information. A set of target data samples is then screened from the initial set of data samples based on the preset redemption policy and the holding policy, where the set of target data samples includes target data samples for a plurality of object samples, and each target data sample satisfies at least one of the redemption policy or the holding policy. Based on the target data sample set, determining the target data tolerance of each object sample through a redemption strategy and a bin holding strategy, taking the target data sample set as the input of a data tolerance prediction model to be trained, outputting the predicted data tolerance of each object sample through the data tolerance prediction model to be trained, and finally training the data tolerance prediction model to be trained based on the target data tolerance of each object sample and the predicted data tolerance of each object sample, so as to obtain the data tolerance prediction model.

In the object tolerance prediction model training stage A2, the predicted data tolerance of each object sample and the object tolerance of each object sample are used as the input of an object tolerance prediction model to be trained, and the predicted object tolerance of each object sample is output through the object tolerance prediction model to be trained, so that the object tolerance prediction model to be trained is trained based on the object tolerance of each object sample and the predicted object tolerance of each object sample, so as to obtain the object tolerance prediction model.

In the object tolerance prediction A3, when the object tolerance of the object to be predicted is required to be predicted in a second time period, first object data of the object to be predicted in a first time period is acquired, the first object data comprises the first object tolerance, at least one of first object access information, first object holding information and first object transaction information, the first time period is the last time period adjacent to the second time period, and then the first object data is subjected to characteristic processing to determine first data characteristics of the object to be predicted. The data tolerance prediction model can be obtained through training in the data tolerance prediction model training stage A1, so that the first data feature of the first object data is used as the input of the data tolerance prediction model, and the first prediction data tolerance of the object to be predicted is output through the data tolerance prediction model. Secondly, in the training stage A2 of the object tolerance prediction model, an object tolerance prediction model can be obtained through training, so that the first prediction data tolerance and the first object tolerance of the object to be predicted are used as the input of the object tolerance prediction model, and the first prediction object tolerance of the object to be predicted in the second time period is output through the object tolerance prediction model, so that the prediction of the object tolerance of the object to be predicted in the second time period is completed.

Further, in the object tolerance prediction A3, a market information set including market information in a plurality of different periods may also be acquired, so that a market prediction influence degree of each object sample in the different periods, which is indicative of influence of the market on behavior of the object sample in one time period, is determined based on the market information in each different period and the target data samples of the plurality of object samples. Therefore, in the foregoing step, the predicted data tolerance of each object sample, the object tolerance of each object sample, and the market prediction influence of each object sample in different periods may also be used as the object tolerance prediction model to be trained, and the predicted object tolerance of each object sample may be output through the object tolerance prediction model to be trained. It should be understood that when the object tolerance and the object basic information of the object sample in each period change with the information such as the object access information, the object holding information and the object transaction information, updated data characteristics can be obtained through the updated information of the target data sample, so that model parameters are continuously adjusted to ensure the accuracy and instantaneity of the model, and the dynamic identification of the data tolerance prediction is realized.

With reference to the foregoing description, taking an execution body as a server as an example, referring to fig. 3, fig. 3 is a schematic diagram of an embodiment of a data processing method provided in an embodiment of the present application, and as shown in fig. 3, the method includes:

101. and acquiring first object data of an object to be predicted in a first time period.

In this embodiment, the server obtains first object data of the object to be predicted in the first time period, where the first object data includes a first object tolerance, and the first object data may further include first object information, where the first object tolerance is used to represent a tolerance degree of a decrease in a resource amount corresponding to a product owned by the object to be predicted in the first time period, where the product owned by the object is an investment product owned by the object, and the resource amount is a holding amount (principal) of the investment product, so that the decrease in the resource amount is a principal loss, and thus the first object tolerance is a principal loss degree acceptable for purchasing the investment product by the object to be predicted in the first time period, for example, principal for purchasing a certain investment product by the object to be predicted is 1000, and a principal loss 150 in the first time period is a principal loss of 15% (150/1000×100%) of the first object tolerance of the object. And the first object information is related behavior information of the object to be predicted on the product in a first time period, and specifically, the first object information comprises at least one of first object access information, first object warehouse holding information and first object transaction information. The first time period may be one week, one month, one quarter or half year, and the specific time period division needs to be flexibly determined according to the application scenario and the actual requirement.

Specifically, the object access information is information corresponding to the object to be predicted accessing one or more investment products or information related to the investment products in a time period, so the object access information may include, but is not limited to, the total number of the investment products accessed by the object to be predicted in the time period, the number of times of accessing any investment product detail page, the number of times of accessing any detail page holding the investment product, the number of times of exposing any investment product detail page, the number of times of exposing different risk grades of the investment product, the number of times of exposing the detail page of the investment product and the holding page of the investment product with different risk grades of the risk grade of the investment product, the maximum value of accessing all the maximum withdrawal of the investment products, the minimum value of accessing all the maximum withdrawal of the investment products, the average value of accessing all the maximum withdrawal of the investment products, and the like.

Secondly, the object bin holding information is information corresponding to bin holding behaviors of the object to be predicted on one or more investment products in a time period, and the bin holding behaviors can be the behaviors of the object to be predicted on a certain investment product from the beginning of the time period to the end of the time period, or the behaviors of the object to be predicted buying a certain investment product in the time period, but not selling the investment product until the end of the time period. Based on this, the object holding information may include, but is not limited to, the maximum withdrawal experienced by each held investment product, the maximum of the maximum withdrawals experienced by all held investment products, the minimum of the maximum withdrawals experienced by all held investment products, the average of the maximum withdrawals experienced by all held investment products, the weighted average of the amounts of the maximum withdrawals experienced by all held investment products, the weighted average of the times of the maximum withdrawals experienced by all held investment products, the held amount duty ratio of the investment products holding different risk levels, etc., for the object to be predicted over the period of time.

Again, the subject transaction information is information corresponding to the subject to be predicted having transacted behavior with respect to one or more investment products during a time period, and may include, but is not limited to, the number of times any investment product is purchased, the number of investment products are purchased, the average value of maximum withdrawals of the investment products to be purchased, the trading number ratio of the investment products to be purchased at different risk levels, the sum ratio of the investment products to be purchased at different risk levels, the maximum value of maximum withdrawals of all the investment products to be purchased, the average value of maximum withdrawals of all the investment products to be purchased, etc. Wherein, the buying is calculated according to the transaction behavior of the object to be predicted on the determined investment product.

Illustratively, if the object to be predicted purchases 100 yuan for investment product a and 100 yuan for investment product B within a time period, investment product a is retracted to 10% within the time period and investment product a is retracted to 2% within the time period. Based on this, the maximum withdrawal of all the investments to be predicted for the subject during the time period is 10%, the minimum withdrawal of all the investments to be predicted during the time period is 2%, and the average value of the withdrawals of all the investments to be predicted during the time period is 6%. The calculation of the maximum value, the minimum value, the average value and other numerical values in the object bin holding information and the object access information is the same, and the details are not repeated here.

Based on this, since the first object data includes a first object tolerance, and the first object data may further include at least one of first object access information, first object holding information, and first object transaction information, the first object data may include the first object tolerance and the first object access information, or the first object tolerance and the first object holding information, or the first object tolerance and the first object transaction information, or the first object tolerance, the first object access information, and the first object holding information, or the first object tolerance, the first object access information, and the first object transaction information, or the first object tolerance, the first object holding information, and the first object transaction information, or the first object tolerance, the first object access information, the first object holding information, and the first object transaction information. It should be appreciated that in practical applications, if any investment product is not accessed, purchased or redeemed for a first period of time, and is not taken out of the warehouse, then the first object data may only include the first object tolerance, and the specific first object data needs to be flexibly determined according to the operation of the investment product by the object to be predicted for the first period of time, so the specific first object data is not limited herein.

102. And carrying out characterization processing on the first object data to determine the first data characteristics of the object to be predicted.

In this embodiment, the server performs a characterization process on the first object data, and determines a first data feature of the object to be predicted. Specifically, since the first object data can include the first object tolerance, and the first object data may further include at least one of the first object access information, the first object holding information and the first object transaction information, it should be understood that, first, if the first object data includes only the first object tolerance, step 102 will not be performed, so in this embodiment, a case where the first object data includes at least one of the first object access information, the first object holding information and the first object transaction information is specifically described, and therefore, performing a characterizing process on the object access information in the first object data can obtain the first access feature, and performing a characterizing process on the object holding information in the first object data can obtain the first holding feature, and performing a characterizing process on the object transaction information in the first object data can obtain the first purchasing feature, so that the first data feature includes at least one of the first access feature, the first purchasing feature and the first holding feature, and the specific feature needs to be determined according to the information included in the first object data, which is not limited.

103. And acquiring the first predicted data tolerance of the object to be predicted through a data tolerance prediction model based on the first data characteristic of the first object data.

In this embodiment, the server uses the first data feature of the first object data as input of the data tolerance prediction model, and outputs the first predicted data tolerance of the object to be predicted through the data tolerance prediction model, where the first predicted data tolerance is the predicted data tolerance of the object to be predicted in the next time period of the first time period. For example, if the first predicted data tolerance of the object to be predicted a is-10%, the predicted data tolerance of the object to be predicted a in the next time period of the first time period is-10%, and similarly, if the first predicted data tolerance of the object to be predicted B is-2%, the predicted data tolerance of the object to be predicted B in the next time period of the first time period is-2%. Secondly, if the first prediction data tolerance of the object to be predicted C is-4%, it is indicated that the prediction data tolerance of the object to be predicted C in the next time period of the first time period is-4%, that is, the prediction data tolerance of the object to be predicted C in the next time period of the first time period is-4%. It should be appreciated that the specific value of the first predicted data tolerance requires a specific first data characteristic determination of the first object data, and thus the foregoing examples should not be construed as limiting the present solution.

104. And acquiring the first predicted object tolerance of the object to be predicted in the second time period through the object tolerance prediction model based on the first predicted data tolerance and the first object tolerance of the object to be predicted.

In this embodiment, the server inputs the first predicted data tolerance of the object to be predicted obtained in step 103 and the first object tolerance obtained in step 102 as an object tolerance prediction model, and outputs the first predicted object tolerance of the object to be predicted in a second time period through the object tolerance prediction model, where the first predicted object tolerance is used to characterize a prediction tolerance degree of a decrease in a resource amount corresponding to a product owned by the object to be predicted in the second time period. The second time period is the next time period adjacent to the first time period, for example, the first time period is 1 month, then the second time period is 2 months, and if the first time period is the first quarter of a year, then the second time period is the second quarter of a year. Second, the first predicted object tolerance is the amount of principal loss that the object to be predicted can accept to purchase the investment product in the second time period. For example, if the tolerance of the first predicted object of the object to be predicted a is-10%, the cost loss level that the object to be predicted a can accept to purchase the investment product in the second time period is-10%, and similarly, if the tolerance of the first predicted object of the object to be predicted B is-4%, the cost loss level that the object to be predicted B can accept to purchase the investment product in the second time period is-4%. It should be appreciated that the specific values of the first predicted object tolerance require flexible determination of the first predicted data tolerance as well as the first object tolerance, and thus the foregoing examples should not be construed as limiting the present solution.

In the embodiment of the application, a data processing method is provided, by adopting the method, the data characteristics of the object to be predicted are constructed through the access information, the bin holding information and other information of multiple dimensions related to the object to be predicted in a time period, the predicted data tolerance is obtained through a data tolerance prediction model based on the constructed data characteristics, and the accuracy of the predicted data tolerance is ensured. And further, based on the predicted data tolerance and the object tolerance with higher accuracy, the predicted object tolerance of the object to be predicted in the next time period is obtained through the object tolerance prediction model, and the problem of low accuracy caused by the behavior deviation of filling the questionnaire by the object is avoided on the basis of considering the object tolerance of the object to be predicted, so that the accuracy of object tolerance prediction on the object is improved.

Optionally, on the basis of the embodiment corresponding to fig. 3, in an optional embodiment of the method for processing data provided in the embodiment of the present application, the method for processing data further includes:

acquiring market information in a first time period;

determining a market prediction influence degree of the object to be predicted based on the market information and the first object data in the first time period, wherein the market prediction influence degree indicates the influence of the trading market on the behavior of the object to be predicted in the first time period;

Based on the first predicted data tolerance and the first object tolerance of the object to be predicted, acquiring the first predicted object tolerance of the object to be predicted in a second time period through an object tolerance prediction model, wherein the method specifically comprises the following steps:

and acquiring the first predicted object tolerance of the object to be predicted in a second time period through the object tolerance prediction model based on the first predicted data tolerance of the object to be predicted, the first object tolerance and the market prediction influence degree of the object to be predicted.

In this embodiment, the server may also obtain market information in the first time period, where the market information may, but is not limited to, prove that the index is an expansion value/a drop value in the first time period, deeply prove that the index is an expansion value/a drop value in the first time period, and that the startup board is an expansion value/a drop value in the first time period. Based on this, the server is able to determine, based on the market information in the first time period and the first object data, a market prediction influence degree of the object to be predicted, which market prediction influence degree indicates an influence of the trading market on the behavior of the object to be predicted in the first time period, i.e. a trading market sensitivity of the object to be predicted. Therefore, the server can further take the first prediction data tolerance of the object to be predicted, the first object tolerance and the market prediction influence of the object to be predicted as inputs of the object tolerance prediction model, and the first prediction object tolerance of the object to be predicted in the second time period is output through the object tolerance prediction model.

For example, if the falling amplitude of the trading market in the first time period is-10%, the investment products held by the objects to be predicted in the first object holding information in the first object data in the first time period are not sold, that is, the falling amplitude of the trading market has a smaller influence on the objects to be predicted. And secondly, if the falling amplitude of the trading market in the first time period is-8%, the investment products held by the objects to be predicted in the first object holding information in the first object data in the first time period are sold for a part, namely the falling amplitude of the trading market is greatly influenced by the objects to be predicted. Or when the falling amplitude in the first time period is-20%, the first object transaction information in the first object data in the first time period indicates that the number of transactions of the object to be predicted in the first time period exceeds the threshold value of the number of transactions (for example, 10), and then the influence degree can be represented by calculating the correlation coefficient through the number of transactions in one time period and the rising and falling amplitude of the duration. It should be understood that the specific market prediction influence of each object to be predicted needs to be determined according to the actual situation, and there may be situations where the market prediction influence of the same object to be predicted in different time periods is different, which will not be further described herein.

For easy understanding, referring to fig. 4, fig. 4 is a schematic flow chart of object tolerance prediction based on market information provided in the embodiment of the present application, and as shown in fig. 4, a server obtains market information B1 in a first time period and obtains first object data B2 of an object to be predicted in the first time period, so as to determine market prediction influence B3 of the object to be predicted based on the market information B1 and the first object data B2 in the first time period. Secondly, by adopting a similar method described in the foregoing embodiment, the first data feature B4 of the object to be predicted is determined by performing the characterizing process on the first object data B2, the first data feature B4 is used as the input of the data tolerance prediction model B5, the first predicted data tolerance B6 of the object to be predicted is output through the data tolerance prediction model B5, and the first object tolerance B7 is included in the first object data B2, so that the server uses the first predicted data tolerance B6 of the object to be predicted, the first object tolerance B7 and the market prediction influence B3 of the object to be predicted as the input of the object tolerance prediction model B8, and the first predicted object tolerance B9 of the object to be predicted in the second time period can be output through the object tolerance prediction model B8. It should be understood that the example of fig. 4 is only used to further detail the process of object tolerance prediction based on market information, and should not be construed as limiting the present application.

In the embodiment of the application, another data processing method is provided, by adopting the method, specific market information in a time period is further introduced, and the market prediction influence degree of the object to be predicted is further constructed based on the object data related to the object to be predicted by the market information, so that the influence of the behavior of the object to be predicted by the trading market can be further learned from the market prediction influence degree when the object tolerance prediction model is used for predicting the object tolerance, and more accurate object tolerance is determined from the actual operation and behavior characteristics of the object to be predicted, and the accuracy of object tolerance prediction is further improved.

Optionally, on the basis of the embodiment corresponding to fig. 3, in an optional embodiment of the method for processing data provided in the embodiment of the present application, the first object data further includes object basic information;

the method for determining the first data characteristics of the first object data comprises the steps of:

carrying out characterization processing on the first object data, and determining first data characteristics of an object to be predicted and object image characteristics of the object to be predicted;

based on the data characteristics of the first object data, acquiring the first predicted data tolerance of the object to be predicted through a data tolerance prediction model, wherein the method specifically comprises the following steps:

And acquiring the first prediction data tolerance of the object to be predicted through a data tolerance prediction model based on the first data characteristic of the first object data and the object portrait characteristic of the object to be predicted.

In this embodiment, the first object data further includes object basic information, which may include, but is not limited to, an age of the object to be predicted, a sex of the object to be predicted, an asset of the object to be predicted, a projectable asset of the object to be predicted, and a consumption capability of the object to be predicted, which is not particularly exhaustive herein. Based on this, the server performs the characterizing processing on the first object data, so that the first data feature of the object to be predicted and the object image feature of the object to be predicted described in the foregoing embodiment can be obtained. Therefore, the server can take the first data characteristics of the first object data and the object portrait characteristics of the object to be predicted as the input of the data tolerance prediction model, and output the first predicted data tolerance of the object to be predicted through the data tolerance prediction model.

In order to facilitate understanding, referring to fig. 5, fig. 5 is a schematic flow chart of obtaining a predicted data tolerance based on an object image feature provided in the embodiment of the present application, as shown in fig. 5, first object data C1 of an object to be predicted in a first period of time of a server, where the first object data includes at least one of a first object tolerance and object basic information, first object access information, first object holding information and first object transaction information, so that by performing a characterizing process on the first object data C1 in the foregoing embodiment, a first data feature C2 of the object to be predicted and an object image feature C3 of the object to be predicted can be obtained, and thus the first data feature C2 of the first object data and the object image feature C3 of the object to be predicted are used as inputs of a data tolerance prediction model C4, and the first predicted data tolerance C5 of the object to be predicted is output through the data tolerance prediction model C4. It should be appreciated that the example of fig. 5 is only used to further detail the flow of obtaining predicted data tolerance based on object representation features and should not be construed as limiting the present application. Second, the predicted data tolerance obtained in FIG. 5 can be used in the embodiment described in step 104 as well as in the flow shown in FIG. 4, without limitation.

In the embodiment of the application, another data processing method is provided, by adopting the method, object basic information is further introduced, and object portrait features of the object to be predicted are further constructed based on the object basic information, so that the object tolerance prediction model can further learn information related to the object and investment from the object portrait features when the data tolerance prediction is performed, thereby improving the accuracy of the obtained predicted data tolerance and further improving the accuracy of the subsequent prediction of the object tolerance.

acquiring second object data of an object to be predicted in a second time period, wherein the second object data comprises second object tolerance and second object information;

characterizing the second object data to determine the second data characteristics of the object to be predicted;

acquiring second predicted data tolerance of the object to be predicted through a data tolerance prediction model based on second data characteristics of the second object data;

And acquiring the second predicted object tolerance of the object to be predicted in a third time period through the object tolerance prediction model based on the second predicted data tolerance of the second object data and the second object tolerance, wherein the third time period is the next time period adjacent to the second time period.

In this embodiment, after obtaining the tolerance of the first object to be predicted in the second time period based on the first object data of the object to be predicted in the first time period, the server needs to continuously predict the tolerance of the object in the next time period, so when the second time period is over, the server can obtain the second object data of the object to be predicted in the second time period in real time, where the second object data includes the second object tolerance and the second object information, and is similar to the first object tolerance and the first object information, the second object tolerance is used to characterize the tolerance degree of the object to be predicted, which is the tolerance degree of the object to be predicted for the decrease in the amount of resources corresponding to the product owned by the object to be predicted in the second time period, and the second object information is related behavior information generated by the object to be predicted for the product in the second time period, so that the second object information includes the second object access information, at least one item of the second object holding information and the second object transaction information, specifically the second object tolerance, the second object access information, the second object holding information and the second object holding information, and the first object holding information and the second transaction information in the step 101, which are similar to the first object holding information and the first object access information.

Based on the data, the server performs characterization processing on the second object data, and determines second data characteristics of the object to be predicted. Specifically, similar to the first object data, since the second object data can include the second object tolerance, and the second object data may further include the second object information, and the second object information includes at least one of the second object access information, the second object holding information and the second object transaction information, and in this embodiment, the second object data includes the second object access information, and the second object holding information and the second object transaction information are specifically described, so that the characterizing the object access information in the second object data can obtain the second access feature, the characterizing the object holding information in the second object data can obtain the second holding feature, and the characterizing the object transaction information in the second object data can obtain the second purchase feature, so that the second data feature includes at least one of the second access feature, the second purchase feature and the second holding feature, and the specific feature needs to be determined according to the information included in the second object data, which is not limited herein.

Further, the server takes the second data characteristic of the second object data as the input of a data tolerance prediction model, and outputs the second predicted data tolerance of the object to be predicted through the data tolerance prediction model, wherein the second predicted data tolerance is the predicted data tolerance of the object to be predicted in the next time period of the second time period. Based on the above, the second predicted data tolerance and the second object tolerance of the object to be predicted obtained through the above steps are input as an object tolerance prediction model, and the second predicted object tolerance of the object to be predicted in a third time period is output through the object tolerance prediction model, where the second predicted object tolerance is used for representing the prediction tolerance degree of the object to be predicted that the corresponding resource amount of the product owned by the object to be predicted is reduced in the third time period. The third time period is the next time period adjacent to the second time period, for example, the first time period is 1 month, the second time period is 2 months, i.e., the third time period is 3 months. Similarly, if the first time period is the first quarter of a year, the second time period is the second quarter of a year, and the third time period is the third quarter of a year. Second, the second predicted object tolerance is the amount of principal loss that the object to be predicted can accept to purchase the investment product in the third time period. The second predicted data tolerance and the second predicted object tolerance are similar to the first predicted data tolerance and the first predicted object tolerance described in step 103 and step 104, and thus are not described herein.

Secondly, the server can also acquire market information in a second time period in a similar manner to that based on the foregoing embodiment, and determine a market prediction influence degree of the object to be predicted based on the market information in the second time period and the second object data, where the market prediction influence degree indicates an influence of the trading market on a behavior of the object to be predicted in the second time period, so that the second prediction object tolerance of the object to be predicted in the second time period is acquired by the object tolerance prediction model based on the second prediction data tolerance of the object to be predicted, the second object tolerance, and the market prediction influence degree of the object to be predicted. And, the second object data can also include object base information of the object to be predicted within the second time period. It should be understood that the specific flow of the object tolerance prediction in the market information and the specific object basic information are similar to those in the foregoing embodiments, and are not repeated here.

In the embodiment of the present application, another data processing method is provided, by adopting the above method, through the risk prediction situation of the previous time period and the information of multiple dimensions of the object to be predicted in the previous time period, the data feature of the object to be predicted is built again, so as to ensure the accuracy of predicting the data tolerance, and further, based on the predicted data tolerance and the object tolerance with higher accuracy, the predicted object tolerance of the object to be predicted in the next time period is obtained through the object tolerance prediction model, thereby realizing dynamic adjustment among multiple periods, ensuring the timeliness of object tolerance prediction, and further improving the accuracy of object tolerance prediction on the object.

Optionally, based on the embodiment corresponding to fig. 3, in an optional embodiment of the method for processing data provided in the embodiment of the present application, the first predicted object tolerance is a second object tolerance.

In this embodiment, the first predicted object tolerance is a second object tolerance, that is, the server directly uses the first predicted object tolerance of the object to be predicted obtained in step 104 in the second time period, that is, the current loss degree that the object to be predicted can accept when purchasing the investment product in the second time period, as the second object tolerance of the second object data of the object to be predicted in the second time period.

In the embodiment of the application, a method for determining the tolerance of a second object is provided, by adopting the method, the tolerance of the predicted object of the next period is obtained through the object data of the previous period, and in the process of carrying out the next period again in the next period, the tolerance of the predicted object is taken as the original tolerance of the object in the next period, so that the object tolerance prediction model can fully obtain the characteristics and the prediction result provided by the object data used in each period, the obtained tolerance of the predicted object is more accurate, and the reliability of prediction among the periods is ensured.

Optionally, on the basis of the embodiment corresponding to fig. 3, in an optional embodiment of the method for processing data provided in the embodiment of the present application, the first object tolerance is determined based on a questionnaire filled in by the object to be predicted in the first time period;

the method for processing data further comprises the following steps:

if the to-be-predicted object updates the questionnaire in the second time period, determining a third object tolerance based on the questionnaire updated by the to-be-predicted object in the second time period;

and if the third object tolerance is smaller than the first predicted object tolerance, determining the third object tolerance as the second object tolerance.

In this embodiment, the first object tolerance is determined based on the questionnaire filled in by the object to be predicted in the first time period, for example, the first object tolerance of the object to be predicted is-10% in the first month of 1 year (in the first time period) based on the questionnaire filled in by the object to be predicted, and then-10% is taken as one item input by the object tolerance prediction model. Or, in the first month of 1 year (in the first time period), the first object tolerance of the object to be predicted is determined to be-15% based on the questionnaire filled in by the object to be predicted, and then-15% is taken as one item of input of the object tolerance prediction model, which is not limited herein.

Based on this, if the object to be predicted does not fill in the questionnaire within the second time period, the server will determine the first predicted object tolerance as the second object tolerance obtained in step 104 based on the method of the previous embodiment. If the object to be predicted updates the questionnaire in the second time period, determining a third object tolerance based on the questionnaire updated by the object to be predicted in the second time period, and determining the third object tolerance as the second object tolerance by the server when the third object tolerance is smaller than the first predicted object tolerance obtained in the step 104.

Illustratively, the first object tolerance of the object to be predicted is determined to be-10% based on the questionnaire filled in by the object to be predicted in the first month (in the first time period) of 1 year, and based on the first predicted object tolerance obtained in the step 104 is-15%, if the questionnaire is updated in the second month (in the second time period) of 1 year, and the determined third object tolerance is-10%, at this time, the third object tolerance-10% is taken as the second object tolerance. Second, if the object to be predicted does not update the questionnaire for the second time period, the first predicted object tolerance is taken as the second object tolerance-15%. Or if the object to be predicted updates the questionnaire within the second time period and the determined tolerance of the third object is-20%, the tolerance of the object will not be immediately adjusted in this embodiment, so the server uses the first predicted object tolerance of-15% as the second object tolerance. It is to be understood that the foregoing examples are provided merely for the understanding of the present disclosure and are not to be construed as limiting the present application.

In the embodiment of the present application, another method for determining the tolerance of the second object is provided, and the tolerance of the predicted object obtained by using the model is determined based on the behavior of the object to be predicted on the questionnaire in the second time period, or the tolerance of the object to be predicted obtained based on the questionnaire is determined, because the questionnaire may have a subjective filling problem of the object to be predicted or a behavior deviation of the object to be predicted in the questionnaire, only when the tolerance of the object to be predicted obtained based on the questionnaire is smaller than the tolerance of the predicted object obtained by using the model, it is described that the risk tolerance of the object to be predicted may be smaller than the previous period, and the tolerance of the object to be predicted based on the questionnaire is used as the tolerance of the object input by using the model in order to avoid investment loss, so that the second object tolerance can accurately reflect the cost loss level that the investment product can be accepted by the object to be predicted in different periods in real time, thereby further improving the flexibility and the timeliness of the object tolerance prediction on the basis of ensuring the accuracy of the object tolerance prediction.

Optionally, on the basis of the embodiment corresponding to fig. 3, in an optional embodiment of the data processing method provided in the embodiment of the present application, the second object information includes second object transaction information, or the second object information includes second object transaction information and second object access information;

The method for processing data further comprises the following steps:

determining a first data tolerance of the object to be predicted within a second period by a redemption policy based on second object transaction information in the second object data;

the first data tolerance is determined as a second object tolerance.

In this embodiment, the second object information includes second object transaction information, or the second object information includes second object transaction information and second object access information, so that it is known that the object performs a transaction operation, for example, a transaction operation including an investment product buy operation or an investment product redemption operation, in the second period. Based on this, the server is also able to determine a first data tolerance of the object to be predicted within the second period by redemption policies based on the second object transaction information included in the second object data, and then the server determines the first data tolerance as the second object tolerance.

Specifically, the redemption strategy includes the situation that the principal amount of the preset investment product exists in the preset investment product special area for purchasing the preset investment product to be predicted, then determining the redemption amount of the preset investment product for the object to be predicted to be redeemed in the preset investment product special area based on the second object transaction information, calculating the difference between the redemption amount of the preset investment product and the principal amount, and calculating the first to-be-selected data tolerance of the preset investment product based on the difference and the principal amount. When a plurality of preset investment products exist in the object to be predicted in one period, the first data tolerance to be selected corresponding to the preset investment products can be obtained, the minimum value (namely the maximum loss) of the first data tolerance to be selected corresponding to the preset investment products is taken as the second data tolerance to be selected, the third data tolerance to be selected is obtained through accumulated calculation of the second data tolerance to be selected based on the first preset percentage, and the data tolerance range between the second data tolerance to be selected and the third data tolerance to be selected can be obtained by the server. Based on this, the server determines the maximum data tolerance of the preset investment product that the object has held in the warehouse in one period based on the second object data, at which time the preset investment product with the maximum data tolerance may not be redeemed, then determines whether the maximum data tolerance of the preset investment product with the maximum data tolerance falls within the data tolerance range between the second to-be-selected data tolerance and the third to-be-selected data tolerance, if yes, the server determines the second to-be-selected data tolerance as the first data tolerance. It will be appreciated that if any of the foregoing redemption policies are not met, the subsequent flow is not triggered and the second object tolerance is determined using the method described in the foregoing embodiments.

Illustratively, taking the first predicted object tolerance obtained by step 104 as-15%, the preset value as 1000, and the first preset percentage as 2% as an example, if the object to be predicted has a purchase preset investment product a, preset investment product B, and preset investment product C principal amount of 1000 in the preset investment product dedicated area, and the redemption amount of the object to be predicted to redeem the preset investment product a in the second time period is 900, the redemption amount to redeem the preset investment product B in the second time period is 920, and the preset investment product C is not redeemed in the second time period, but the minimum holding amount (i.e., the loss is the largest) of 910 in the second time period.

Based on this, the server calculates the difference between the redemption amount of the preset investment product a and the principal amount based on the numerical value of the previous example to be-100 (900-1000= -100), obtains the first data tolerance of the preset investment product a to be-10% [ -100/1000 x 100% ] based on the difference (-100) and the principal amount (1000), and similarly obtains the first data tolerance of the preset investment product B to be-8% [ -80/1000 x 100% ] based on the difference (-80) and the principal amount (1000) to be-80 (980-1000= -80). The first data tolerance (-10%) of the preset investment product a and the first data tolerance (-8%) of the preset investment product B are determined to be smaller in value (i.e., the deficit is the largest) as the second data tolerance.

Therefore, the server will determine that the first data tolerance (-10%) of the preset investment product a is the second data tolerance to be selected, and add the second data tolerance (-10%) to the first preset percentage 2% to obtain a third data tolerance to be selected-8%, thereby obtaining a data tolerance range between the second data tolerance (-10%) and the third data tolerance (-8%) to be selected-8% to-10%. The server further calculates the maximum data tolerance of the preset investment product C at this time, that is, calculates the difference between the minimum data tolerance and the principal sum to be-90 (910-1000), obtains the maximum data tolerance of the preset investment product C to be-9% [ -90/1000 x 100% ] based on the difference (-90) and the principal sum (1000), and determines that the maximum data tolerance (-9%) of the preset investment product C falls within the data tolerance range (-8% to-10%) between the second data tolerance to be selected and the third data tolerance to be selected, so that at this time, the second data tolerance to be selected is determined to be the first loss, and the server further determines the first data tolerance (-10%) to be the second object tolerance.

It is to be understood that the foregoing examples are provided merely for the understanding of the present invention and are not to be construed as limiting thereof.

In the embodiment of the application, another method for determining the tolerance of the second object is provided, and the method is adopted, based on the transaction operation of the object to be predicted in a time period, namely, the redemption operation of the object is adopted to determine which data tolerance the object cannot bear in the period, so that the determination of the data tolerance is regulated through the specific redemption operation of the object, and the flexibility and the timeliness of the prediction of the tolerance of the object are improved.

Optionally, in an optional embodiment of the method for processing data provided in the embodiment of the present application on the basis of the embodiment corresponding to fig. 3, the second object information includes second object holding information, or the second object information includes second object holding information and second object access information;

the method for processing data further comprises the following steps:

determining a second data tolerance of the object to be predicted in a second period based on second object holding information in the second object data;

and if the second data tolerance is smaller than the first predicted object tolerance, determining the second data tolerance as the second object tolerance.

In this embodiment, the second object information includes second object holding information, or the second object information includes second object holding information and second object access information, and since the object holding information is information corresponding to a holding behavior of an object to be predicted for one or more investment products in a time period, the holding behavior may be a behavior that the object to be predicted holds an unsold product from a beginning of the time period to an end of the time period, or a behavior that the object to be predicted purchases an investment product in the time period, but holds an unsold product until the end of the time period, so that the object still has a holding behavior for at least one investment product in the second period. Based on this, the server can also determine, based on the second object holding information in the second object data, a second data tolerance of the object to be predicted in the second period, where the second data tolerance is the smallest data tolerance (i.e., the deficit is largest) in the investment products in which the object to be predicted has a holding behavior in the second period, and when the second data tolerance is smaller than the first predicted object tolerance, the server determines the second data tolerance as the second object tolerance.

Illustratively, taking the first predicted object tolerance obtained in step 104 as-10%, the preset value as 1000, and the second preset percentage as 2% as an example, if the object to be predicted has a value of 1000 for purchasing preset investment product a, preset investment product B, and preset investment product C in the preset investment product dedicated area, and the object to be predicted does not redeem preset investment product a, preset investment product B, and preset investment product C for the second time period, if the minimum holding amount of preset investment product a for the second time period is 850, the minimum holding amount of preset investment product B for the second time period is 880, and the minimum holding amount of preset investment product C for the second time period is 900.

Based on this, the server calculates the difference between the lowest holding amount and principal amount of the preset investment product a to be-150 (850-1000= -150), obtains the fourth to-be-selected data tolerance of the preset investment product a to be-15% [ -150/1000 x 100% ] based on the difference (-150) and principal amount (1000), and similarly, obtains the fourth to-be-selected data tolerance of the preset investment product C to be-10% [ -100/1000 x 100% ] based on the difference (-120) and principal amount (1000), and obtains the fourth to-be-selected data tolerance of the preset investment product B to be-12% [ -120/1000 x 100% ] based on the difference (-100) and principal amount (1000). Therefore, the tolerance of the fourth to-be-selected data of the preset investment product a is (-15%), the tolerance of the fourth to-be-selected data of the preset investment product B is (-12%), and the tolerance of the fourth to-be-selected data of the preset investment product C is (-10%), and the tolerance of the fourth to-be-selected data is the smallest (i.e., the loss is the largest) among the preset investment products a to C, so the server will determine that the tolerance of the fourth to-be-selected data of the preset investment product a (-15%) is the second data tolerance. And since the second data tolerance (-15%) is less than the first predicted object tolerance (-10%), the server determines the second data tolerance (-15%) as the second object tolerance.

In the embodiment of the application, another data processing method is provided, and the method is adopted, so that the maximum data tolerance of the object which is not redeemed in the period is determined based on the object to be predicted in the period, namely, the object is taken as the object, and when the data tolerance is smaller than the predicted object tolerance, the investment product actually taken as the object is not redeemed in the period after reaching the predicted object tolerance, so that the risk tolerance of the object is possibly improved, and the determination of the data tolerance is adjusted through the specific object bin behavior, so that the flexibility and the timeliness of the object tolerance prediction are improved.

Optionally, in an optional embodiment of the method for processing data provided in the embodiment of the present application based on the embodiment corresponding to fig. 3, the data feature includes at least one of an access feature, a purchase feature, and a holding feature, where the access feature is obtained based on the object access information, the purchase feature is obtained based on the object transaction information, and the holding feature is obtained based on the object holding information;

In this embodiment, the data features include at least one of an access feature, a purchase feature, and a holding feature, similar to those mentioned in the previous embodiment, the access feature being based on the object access information, the purchase feature being based on the object transaction information, and the holding feature being based on the object holding information. Thus, the data tolerance prediction model specifically includes an access data tolerance prediction model for obtaining an access prediction data tolerance of the object based on the access characteristic, a purchase application data tolerance prediction model for obtaining a purchase application prediction data tolerance of the object based on the purchase application characteristic, and a holding data tolerance prediction model for obtaining a holding prediction data tolerance of the object based on the holding characteristic.

Based on the above, since the object access information, the object transaction information and the object holding information of different objects to be predicted are different, the corresponding access feature, the purchasing feature and the holding feature included in the obtained data feature are also different, and the server respectively inputs the three different features of the access feature, the purchasing feature and the holding feature into different prediction models in the data tolerance prediction model, that is, if the data feature includes the access feature obtained based on the object access information, the server takes the access feature in the data feature as the input of the access data tolerance prediction model in the data tolerance prediction model, and at this time, the access data tolerance prediction model outputs the access prediction data tolerance of the object to be predicted based on the access feature. Similarly, if the data features include the purchasing features obtained based on the object transaction information, the server takes the purchasing features in the data features as the input of the purchasing data tolerance prediction model in the data tolerance prediction model, and the purchasing data tolerance prediction model outputs the purchasing prediction data tolerance of the object to be predicted based on the purchasing features. And if the data features comprise the holding features obtained based on the object holding information, the server takes the holding features in the data features as the input of a holding data tolerance prediction model in the data tolerance prediction model, and the holding data tolerance prediction model outputs the holding prediction data tolerance of the object to be predicted based on the holding features. If the method includes the above multiple features, the server inputs different features into corresponding data tolerance prediction models in the data tolerance prediction models, so that the obtained first prediction data tolerance of the object to be predicted includes at least one of access prediction data tolerance of the object to be predicted, purchase prediction data tolerance of the object to be predicted and holding prediction data tolerance of the object to be predicted.

For ease of understanding, taking the first object data including the object access information, the object transaction information and the object holding information as an example for explanation, please refer to fig. 6, fig. 6 is a schematic flow chart of the predicted data tolerance determination provided in the embodiment of the present application, D1 refers to the first object data, D2 refers to the first data feature, D3 refers to the data tolerance prediction model, and D4 refers to the first predicted data tolerance. Based on the above, the server performs characteristic processing on different object access information, object transaction information and object holding information in the first object data D1, so as to obtain access characteristics, purchase characteristics and holding characteristics in the first data characteristics D2, then uses the access characteristics, purchase characteristics and holding characteristics in the data characteristics D2 as the access data tolerance prediction model in the data tolerance prediction model D3, and inputs the access prediction data tolerance by the access data tolerance prediction model in the data tolerance prediction model D3, outputs the access prediction data tolerance by the purchase data tolerance prediction model in the data tolerance prediction model D3, outputs the holding prediction data tolerance by the holding data tolerance prediction model in the data tolerance prediction model D3, and outputs the holding prediction data tolerance by the holding data tolerance prediction model in the data tolerance prediction model D3, so as to obtain the first prediction data including the access prediction data tolerance of the object to be predicted, the purchase prediction data tolerance of the object to be predicted and the holding prediction data tolerance of the object to be predicted. It should be understood that, in practical applications, since specific information included in the first object data is different, the obtained data features are also different, and further, the processing of subsequent steps and the obtained defect value types are also different, so that the foregoing examples are only used for understanding the present solution, and the specific predicted data tolerance and the applied data tolerance prediction model need to be flexibly determined according to practical situations.

In the embodiment of the application, a method for determining the tolerance of predicted data is provided, by adopting the method, based on information of different dimensions included in first object data, different characterization processing is performed to obtain data characteristics respectively corresponding to the information of different dimensions, so that the flexibility of the scheme is improved. Secondly, different data features acquire corresponding predicted data tolerance through corresponding data tolerance prediction models, so that the obtained predicted data tolerance comprises predicted values under a plurality of dimension information, and more accurate feature processing is guaranteed on the features corresponding to each dimension through the corresponding data tolerance prediction models, so that the determined predicted data tolerance is more accurate.

The method for performing data processing in practical application described in the above embodiment will be described below as a method for performing model training before performing object tolerance prediction to obtain a data tolerance prediction model and an object tolerance prediction model.

Acquiring an initial data sample set, wherein the initial data sample set comprises initial data samples of a plurality of object samples, and each initial data sample comprises at least one of object tolerance, object basic information, object access information, object warehouse holding information and object transaction information;

screening a target data sample set from the initial data sample set based on the redemption policy and the holding policy, wherein the target data sample set includes target data samples of the plurality of object samples;

determining a target data tolerance for each subject sample by redemption policies and holding policies based on the target data sample set;

based on the target data sample set, obtaining the predicted data tolerance of each object sample through a data tolerance prediction model to be trained;

based on the target data tolerance of each object sample and the predicted data tolerance of each object sample, training the data tolerance prediction model to be trained to obtain the data tolerance prediction model.

In this embodiment, the server first obtains an initial data sample set, where the initial data sample set includes initial data samples of a plurality of object samples, and each initial data sample includes at least one of object tolerance, object basic information, object access information, object holding information, and object transaction information. The subject tolerance in the initial data sample is determined from the questionnaire filled in by the subject sample.

The server then screens a set of target data samples from the initial set of data samples based on the redemption policy and the holding policy, the set of target data samples comprising a plurality of target data samples, each target data sample being capable of satisfying either the redemption policy or the holding policy, i.e., each target data sample being capable of determining a data tolerance based on either the redemption policy or the holding policy, and if the redemption policy and the holding policy are not satisfied, indicating that the initial data sample has a data bias or a data drift condition, such data being directly rejected. Thus, the server will determine the target data tolerance for each subject sample based on the set of target data samples, via a redemption policy, similar to that described in the previous embodiments, and thus the manner in which the target data tolerance is determined via the redemption policy is also similar to that described in the previous embodiments, and will not be repeated here. The method of determining data tolerance by the binning strategy will be described below.

Specifically, the aforementioned holding policy includes that the situation that the principal amount of the preset investment product is greater than or equal to the preset value exists in the special purchasing area of the preset investment product for the object to be predicted, then the lowest holding amount (that is, the holding amount when the loss of each held preset investment product is the largest) of all holding preset investment products of the object to be predicted in the special purchasing area of the preset investment product is determined based on the holding information of the second object, the difference between the lowest holding amount of each holding preset investment product and the principal amount of the holding preset investment product is calculated, and the fourth data tolerance to be selected of the preset investment product is calculated based on the difference and the principal amount. When the object to be predicted has a plurality of preset investment products in one period, fourth data tolerance to be selected corresponding to the plurality of preset investment products can be obtained through the steps, and the minimum value (namely the maximum loss) of the fourth data tolerance to be selected corresponding to the plurality of preset investment products is taken as the fifth data tolerance to be selected. Based on the above, the server can also determine the tolerance of the sixth to-be-selected object based on the latest questionnaire filled in by the to-be-predicted object, and perform subtraction processing on the tolerance of the sixth to-be-selected object and the second preset percentage to obtain the tolerance of the seventh to-be-selected data, and determine the tolerance of the fifth to-be-selected data as the target data tolerance when the tolerance of the fifth to-be-selected data is greater than the tolerance of the seventh to-be-selected object. It will be appreciated that if any of the aforementioned retention policies is not met, the subsequent flow is not triggered, and the targeted data tolerance is determined using a similar method as described for the redemption policy in the aforementioned embodiments.

Illustratively, consider the example where the tolerance of the first predicted object is-10%, the predetermined value is 1000, the second predetermined percentage is 2%, and the determination of the tolerance of the sixth object to be selected is-15% based on the latest questionnaire filled in by the object to be predicted, if the object to be predicted has a purchase of the predetermined investment product a, the predetermined investment product B, and the predetermined investment product C principal amount of 1000 in the predetermined investment product dedicated area, and the object to be predicted does not redeem the predetermined investment product a, the predetermined investment product B, and the predetermined investment product C, if the minimum holding amount of the predetermined investment product a is 850, the minimum holding amount of the predetermined investment product B is 880, and the minimum holding amount 900 of the predetermined investment product C.

Based on this, the server calculates the difference between the lowest holding amount and principal amount of the preset investment product a to be-150 (850-1000= -150), obtains the fourth to-be-selected data tolerance of the preset investment product a to be-15% [ -150/1000 x 100% ] based on the difference (-150) and principal amount (1000), and similarly, obtains the fourth to-be-selected data tolerance of the preset investment product C to be-10% [ -100/1000 x 100% ] based on the difference (-120) and principal amount (1000), and obtains the fourth to-be-selected data tolerance of the preset investment product B to be-12% [ -120/1000 x 100% ] based on the difference (-100) and principal amount (1000).

Therefore, the tolerance of the fourth to-be-selected data of the preset investment product a is (-15%), the tolerance of the fourth to-be-selected data of the preset investment product B is (-12%), and the tolerance of the fourth to-be-selected data of the preset investment product C is (-10%), and the tolerance of the fourth to-be-selected data is the smallest (i.e., the loss is the largest) among the preset investment products a to C, so the server will determine that the tolerance of the fourth to-be-selected data of the preset investment product a (-15%) is the tolerance of the fifth to-be-selected data. Secondly, adding the tolerance (-15%) of the sixth to-be-selected object to the second preset percentage (2%) to obtain a tolerance (-13%) of the seventh to-be-selected data, wherein the tolerance (-15%) of the fifth to-be-selected data is smaller than the tolerance (-13%) of the seventh to-be-selected object, and the true tolerance that the specification object can accept is more, so that the tolerance (-15%) of the fifth to-be-selected data needs to be determined as the target data tolerance, and the accuracy of the determined target data tolerance can be improved. It is to be understood that the foregoing examples are provided merely for the understanding of the present invention and are not to be construed as limiting thereof.

Further, the server inputs the target data sample set as a data tolerance prediction model to be trained, outputs the predicted data tolerance of each object sample through the data tolerance prediction model to be trained, and trains the data tolerance prediction model to be trained based on the target data tolerance of each object sample and the predicted data tolerance of each object sample so as to obtain the data tolerance prediction model.

Specifically, the server performs iterative training with the target data tolerance of each object sample as a target, namely, determines a loss value of the loss function according to the difference between the target data tolerance of each object sample and the predicted data tolerance of each object sample, judges whether the loss function reaches a convergence condition according to the loss value of the loss function, and if the convergence condition is not reached, updates model parameters of a data tolerance prediction model to be trained by using the loss value of the loss function until the loss function reaches the convergence condition, and determines the loss value of the loss function as the model parameters of the data tolerance prediction model to be trained, thereby obtaining the data tolerance prediction model.

In order to understand the foregoing training process, fig. 7 is a schematic flow chart of training a data tolerance prediction model to be trained according to an embodiment of the present application, where E1 refers to an initial data sample set, E2 refers to a target data sample set, E3 refers to a target data tolerance of each object sample, E4 refers to a data tolerance prediction model to be trained, and E5 refers to a predicted data tolerance of each object sample. Based on this, the server screens the initial data sample set E1 in the manner described in the foregoing embodiment to obtain the target data sample set E2, then determines the target data tolerance E3 of each object sample based on the target data sample set E2, inputs the target data sample set E2 into the data tolerance prediction model E4 to be trained, obtains the predicted data tolerance E5 of each object sample through the data tolerance prediction model E4 to be trained, and then performs iterative training on the data tolerance prediction model E4 to be trained based on the target data tolerance E3 of each object sample, the predicted data tolerance E5 of each object sample and the loss function, that is, performs updating of the model parameters of the data tolerance prediction model E4 to be trained until the updated model parameters make the loss function reach the convergence condition, so as to generate the data tolerance prediction model. It should be understood that the example in fig. 7 is merely for convenience in understanding the present scheme, and is not intended to limit the present scheme.

Secondly, as can be seen from the foregoing embodiments, since the data tolerance prediction model includes the visit data tolerance prediction model, the purchase data tolerance prediction model and the holding data tolerance prediction model, in the actual training process, the to-be-trained data tolerance prediction model includes the visit data tolerance prediction model, the purchase data tolerance prediction model and the holding data tolerance prediction model, and therefore the visit data tolerance prediction model, the purchase data tolerance prediction model and the holding data tolerance prediction model need to be trained respectively.

Similarly to the previous steps, the server takes the object access information sample in the object data of each object sample in the object data sample set as the input of the access data tolerance prediction model to be trained, the access data tolerance prediction model to be trained outputs the access prediction data tolerance of each object sample, and accordingly, iterative training is conducted by taking the object data tolerance of each object sample as a target, namely, the loss value of the loss function is determined according to the difference between the object data tolerance of each object sample and the access prediction data tolerance of each object sample, whether the loss function reaches the convergence condition is judged according to the loss value of the loss function, if the convergence condition is not reached, the model parameters of the access data tolerance prediction model to be trained are updated by using the loss value of the loss function, until the loss function reaches the convergence condition, the loss value of the loss function is determined as the model parameters of the access data tolerance prediction model to be trained, and the access data tolerance prediction model is obtained.

Similarly, the server takes object transaction information samples in object data of each object sample in the object data sample set as input of a to-be-trained purchasing data tolerance prediction model, the to-be-trained purchasing data tolerance prediction model outputs purchasing prediction data tolerance of each object sample, so that the object data tolerance of each object sample is taken as a target to carry out iterative training, namely, a loss value of a loss function is determined according to the difference between the object data tolerance of each object sample and the purchasing prediction data tolerance of each object sample, whether the loss function reaches a convergence condition is judged according to the loss value of the loss function, if the convergence condition is not reached, model parameters of the to-be-trained purchasing data tolerance prediction model are updated by the loss value of the loss function until the loss function reaches the convergence condition, and the loss value of the loss function is determined as model parameters of the to-be-trained purchasing data tolerance prediction model, so that the purchasing data tolerance prediction model is obtained.

And the server takes an object holding information sample in object data of each object sample in the object data sample set as input of a holding data tolerance prediction model to be trained, the holding data tolerance prediction model to be trained outputs the holding prediction data tolerance of each object sample, so that the object data tolerance of each object sample is used as a target for iterative training, namely, a loss value of a loss function is determined according to the target data tolerance of each object sample and the difference between the holding prediction data tolerances of each object sample, whether the loss function reaches a convergence condition is judged according to the loss value of the loss function, if the convergence condition is not reached, the loss value of the loss function is used for updating model parameters of the holding data tolerance prediction model to be trained, until the loss function reaches the convergence condition, the loss value of the loss function is determined as model parameters of the holding data tolerance prediction model to be trained, and the holding data tolerance prediction model is obtained. The specific training process of the access data tolerance prediction model to be trained, the purchase data tolerance prediction model to be trained and the holding data tolerance prediction model to be trained is similar to that of fig. 7, and will not be repeated here.

It should be understood that the convergence condition of the foregoing loss function may be that the value of the loss function is less than or equal to the first preset threshold, for example, the value of the first preset threshold may be 0.005, 0.01, 0.02 or other values approaching 0; the difference between the values of two adjacent times of the loss function may be less than or equal to a second preset threshold, where the value of the second threshold may be the same as or different from the value of the first threshold, for example, the value of the second preset threshold may be 0.005, 0.01, 0.02, or other values approaching 0, and the server may also use other convergence conditions, and the server is not limited herein.

According to the method for training the data tolerance prediction model, the initial data samples can be screened before model training, the data set obtained by model training is accurate and data are not cheap, and therefore the efficiency and reliability of model training are guaranteed. Secondly, different types of data tolerance prediction models to be trained in the data tolerance prediction models can be trained respectively, so that different data tolerance prediction models which can obtain corresponding predicted data tolerance based on different data characteristics are obtained, more accurate characteristic processing is guaranteed to the characteristics corresponding to each dimension through the corresponding data tolerance prediction models, and therefore flexibility of model training is improved.

based on the predicted data tolerance of each object sample and the object tolerance of each object sample, obtaining the predicted object tolerance of each object sample through an object tolerance prediction model to be trained;

based on the object tolerance of each object sample and the predicted object tolerance of each object sample, training the object tolerance prediction model to be trained to obtain an object tolerance prediction model.

In this embodiment, the server may also use the predicted object tolerance of each object sample obtained in the foregoing embodiment and the object tolerance of each object sample as input of the object tolerance prediction model to be trained, output the predicted object tolerance of each object sample through the object tolerance prediction model to be trained, and train the object tolerance prediction model to be trained based on the object tolerance of each object sample and the predicted object tolerance of each object sample, so as to obtain the object tolerance prediction model. Specifically, the server performs iterative training with the object tolerance of each object sample as a target, namely, determines a loss value of the loss function according to the difference between the object tolerance of each object sample and the predicted object tolerance of each object sample, judges whether the loss function reaches a convergence condition according to the loss value of the loss function, and if the convergence condition is not reached, updates model parameters of the object tolerance prediction model to be trained by using the loss value of the loss function until the loss function reaches the convergence condition, and determines the loss value of the loss function as the model parameters of the object tolerance prediction model to be trained, thereby obtaining the object tolerance prediction model.

In order to understand the foregoing training process, fig. 8 is a schematic flow chart of training a data tolerance prediction model to be trained according to an embodiment of the present application, where F1 refers to a predicted data tolerance of each object sample, F2 refers to an object tolerance of each object sample, F3 refers to a target tolerance prediction model to be trained, and F4 refers to a predicted object tolerance of each object sample. Based on the above, the server inputs the predicted data tolerance F1 of each object sample and the object tolerance F2 of each object sample into the object tolerance prediction model F3 to be trained, obtains the predicted object tolerance F4 of each object sample through the object tolerance prediction model F3 to be trained, and then carries out iterative training on the object tolerance prediction model F3 to be trained based on the object tolerance F2 of each object sample, the predicted object tolerance F4 of each object sample and the loss function, namely carries out updating of model parameters of the object tolerance prediction model F3 to be trained until the updated model parameters enable the loss function to reach convergence conditions, so as to generate the object tolerance prediction model. It should be understood that the example in fig. 8 is merely for convenience in understanding the present scheme, and is not intended to limit the present scheme.

In the embodiment of the application, another data processing method is provided, and in the model training, the predicted object tolerance is predicted based on the predicted data tolerance of the object sample output by the previous model and the object tolerance of the object sample, and thus model training is performed based on the object tolerance and the predicted object tolerance, so that the model learns more characteristic information in the training process, and the reliability of model training is ensured.

Acquiring a market information set, wherein the market information set comprises market information in a plurality of different periods;

determining a market prediction influence degree of each object sample in different periods based on market information in each different period and target data samples of a plurality of object samples, wherein the market prediction influence degree indicates the influence of the market on the behaviors of the object samples in one time period;

based on the predicted data tolerance of each object sample and the object tolerance of each object sample, obtaining the predicted object tolerance of each object sample through an object tolerance prediction model to be trained, wherein the predicted object tolerance comprises the following steps:

and obtaining the predicted object tolerance of each object sample through the object tolerance prediction model to be trained based on the predicted data tolerance of each object sample, the object tolerance of each object sample and the market prediction influence degree of each object sample in different periods.

In this embodiment, a method for model training by adding a market information set is provided. Therefore, the server may further obtain a market information set, where the market information set includes market information in a plurality of different periods, and then determine, based on the market information in each different period and the target data samples of the plurality of object samples, a market prediction influence degree of each object sample in the different periods, where the market prediction influence degree indicates an influence of the market on the behavior of the object sample in one time period, and a specific determination manner is similar to that of the foregoing embodiment, and is not repeated herein. Based on the above, the server takes the predicted data tolerance of each object sample, the object tolerance of each object sample and the market prediction influence degree of each object sample in different periods as the input of the object tolerance prediction model to be trained, and outputs the predicted object tolerance of each object sample through the object tolerance prediction model to be trained.

According to the method, specific market information in a plurality of time periods is further introduced in model training, and the market prediction influence degree of the object sample is further built based on object data related to the market information and the object sample, so that the object tolerance prediction model can further learn influence of a transaction market on behaviors of the object sample from the market prediction influence degree during training, and more accurate object tolerance is determined from actual operation and behavior characteristics of an object to be predicted, and reliability of model training is further improved.

Fig. 9 is a schematic structural diagram of a data processing apparatus according to an embodiment of the present application, as shown in fig. 9, the data processing apparatus 900 includes:

an obtaining module 901, configured to obtain first object data of an object to be predicted in a first time period, where the first object data includes a first object tolerance and first object information;

a determining module 902, configured to perform a characterization process on the first object data, and determine a first data feature of the object to be predicted;

the obtaining module 901 is further configured to obtain a first predicted data tolerance of the object to be predicted through a data tolerance prediction model based on a first data feature of the first object data;

The obtaining module 901 is further configured to obtain, based on the first predicted data tolerance of the object to be predicted and the first object tolerance, a first predicted object tolerance of the object to be predicted within a second time period, where the second time period is a next time period adjacent to the first time period, by using an object tolerance prediction model.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, the obtaining module 901 is further configured to obtain market information in a first time period;

the determining module 902 is further configured to determine a market prediction influence degree of the object to be predicted based on the market information and the first object data in the first time period, where the market prediction influence degree indicates an influence of the trading market on a behavior of the object to be predicted in the first time period;

the obtaining module 901 is specifically configured to obtain, by using an object tolerance prediction model, a first predicted object tolerance of an object to be predicted in a second time period based on a first predicted data tolerance of the object to be predicted, the first object tolerance, and a market prediction influence of the object to be predicted.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, the first object data further includes object basic information;

the determining module 902 is specifically configured to determine a first data feature of the object to be predicted and an object image feature of the object to be predicted based on characterizing the first object data;

the obtaining module 901 is specifically configured to obtain, based on a first data feature of the first object data and an object representation feature of an object to be predicted, a first predicted data tolerance of the object to be predicted through a data tolerance prediction model.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, the obtaining module 901 is further configured to obtain second object data of an object to be predicted in a second time period, where the second object data includes a second object tolerance and second object information;

the determining module 902 is further configured to perform a characterization process on the second object data, and determine a second data feature of the object to be predicted;

the obtaining module 901 is further configured to obtain a second predicted data tolerance of the object to be predicted through the data tolerance prediction model based on a second data feature of the second object data;

The obtaining module 901 is further configured to obtain, based on a second predicted data tolerance of the second object data and the second object tolerance, a second predicted object tolerance of the object to be predicted within a third time period, where the third time period is a next time period adjacent to the second time period, by using an object tolerance prediction model.

Optionally, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, based on the embodiment corresponding to fig. 9, the first predicted object tolerance is a second object tolerance.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, the first object tolerance is determined based on a questionnaire filled in by the object to be predicted in the first time period;

the determining module 902 is further configured to determine a third object tolerance based on the questionnaire updated by the object to be predicted in the second time period if the object to be predicted updates the questionnaire in the second time period;

the determining module 902 is further configured to determine the third object tolerance as the second object tolerance if the third object tolerance is less than the first predicted object tolerance.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, the second object information includes second object transaction information, or the second object information includes second object transaction information and second object access information;

a determining module 902 further configured to determine, based on second object transaction information in the second object information, a first data tolerance of the object to be predicted in a second period through a redemption policy;

the determining module 902 is further configured to determine the first data tolerance as a second object tolerance.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, the second object information includes second object holding information, or the second object information includes second object holding information and second object access information;

the determining module 902 is further configured to determine a second data tolerance of the object to be predicted in a second period based on second object binning information in the second object information;

the determining module 902 is further configured to determine the second data tolerance as the second object tolerance if the second data tolerance is less than the first predicted object tolerance.

Optionally, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, on the basis of the embodiment corresponding to fig. 9, the data features include at least one of an access feature, a purchase feature, and a holding feature, where the access feature is obtained based on the object access information, the purchase feature is obtained based on the object transaction information, and the holding feature is obtained based on the object holding information;

Optionally, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, based on the embodiment corresponding to fig. 9, the data processing apparatus 900 further includes a screening module 903 and a training module 904;

the acquiring module 901 is further configured to acquire an initial data sample set, where the initial data sample set includes initial data samples of a plurality of object samples, and each initial data sample includes at least one of object tolerance and object basic information, object access information, object holding information, and object transaction information;

a screening module 903 for screening a set of target data samples from the initial set of data samples based on the redemption policy and the holding policy, wherein the set of target data samples includes target data samples for a plurality of object samples;

a determination module 902 further for determining a target data tolerance for each of the object samples based on the set of target data samples by redemption policies and the holding policies;

the obtaining module 901 is further configured to obtain a predicted data tolerance of each object sample through a data tolerance prediction model to be trained based on the target data sample set;

the training module 904 is configured to train the data tolerance prediction model to be trained based on the target data tolerance of each object sample and the predicted data tolerance of each object sample, so as to obtain the data tolerance prediction model.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in this embodiment of the present application, the obtaining module 901 is further configured to obtain, based on the predicted data tolerance of each object sample and the object tolerance of each object sample, the predicted object tolerance of each object sample through an object tolerance prediction model to be trained;

the training module 904 is further configured to train the object tolerance prediction model to be trained based on the object tolerance of each object sample and the predicted object tolerance of each object sample, so as to obtain the object tolerance prediction model.

Optionally, on the basis of the embodiment corresponding to fig. 9, in another embodiment of the data processing apparatus 900 provided in the embodiment of the present application, the obtaining module 901 is further configured to obtain a market information set, where the market information set includes market information in a plurality of different periods;

a determining module 902, configured to determine a market prediction impact level of each object sample in different periods based on market information in each different period and target data samples of the plurality of object samples, where the market prediction impact level indicates an impact of the market on a behavior of the object sample in one time period;

The obtaining module 901 is specifically configured to obtain, based on the predicted data tolerance of each object sample, the object tolerance of each object sample, and the market prediction influence of each object sample in different periods, the predicted object tolerance of each object sample through an object tolerance prediction model to be trained.

The embodiments of the present application further provide another data processing apparatus, where the data processing apparatus may be disposed on a server, and may also be disposed on a terminal device, and in this application, the data processing apparatus is disposed on a server for explanation, referring to fig. 10, fig. 10 is a schematic diagram of an embodiment of a server in this application, where the server 1000 may, as shown in the drawing, generate relatively large differences due to different configurations or performances, may include one or more central processing units (central processing units, CPU) 1022 (for example, one or more processors) and a memory 1032, and one or more storage media 1030 (for example, one or more mass storage devices) storing application programs 1042 or data 1044. Wherein memory 1032 and storage medium 1030 may be transitory or persistent. The program stored on the storage medium 1030 may include one or more modules (not shown), each of which may include a series of instruction operations on a server. Further, central processor 1022 may be configured to communicate with storage medium 1030 to perform a series of instruction operations in storage medium 1030 on server 1000.

The Server 1000 may also include one or more power supplies 1026, one or more wired or wireless network interfaces 1050, one or more input/output interfaces 1058, and/or one or more operating systems 1041, such as Windows Server ^TM ，Mac OS X ^TM ，Unix ^TM ，Linux ^TM ，FreeBSD ^TM Etc.

The steps performed by the server in the above embodiments may be based on the server structure shown in fig. 10.

The CPU 1022 included in the server is used to execute the embodiments shown in fig. 3 and the respective embodiments corresponding to fig. 3.

The present application further provides a terminal device, as shown in fig. 11, for convenience of explanation, only a portion related to an embodiment of the present application is shown, and specific technical details are not disclosed, please refer to a method portion of an embodiment of the present application. The terminal device is taken as a mobile phone for example for explanation:

fig. 11 is a block diagram showing a part of the structure of a mobile phone related to a terminal provided in an embodiment of the present application. Referring to fig. 11, the mobile phone includes: radio Frequency (RF) circuitry 1110, memory 1120, input unit 1130, display unit 1140, sensors 1150, audio circuit 1160, wireless fidelity (wireless fidelity, wiFi) module 1170, processor 1180, power supply 1190, and the like. Those skilled in the art will appreciate that the handset configuration shown in fig. 11 is not limiting of the handset and may include more or fewer components than shown, or may combine certain components, or may be arranged in a different arrangement of components.

The following describes the components of the mobile phone in detail with reference to fig. 11:

the RF circuit 1110 may be used for receiving and transmitting signals during a message or a call, and in particular, after receiving downlink information of a base station, the downlink information is processed by the processor 1180; in addition, the data of the design uplink is sent to the base station. Typically, the RF circuitry 1110 includes, but is not limited to, an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (Low Noise Amplifier, LNA), a duplexer, and the like. In addition, RF circuitry 1110 may also communicate with networks and other devices via wireless communications. The wireless communications may use any communication standard or protocol including, but not limited to, global system for mobile communications (Global System of Mobile communication, GSM), general packet radio service (General Packet Radio Service, GPRS), code division multiple access (Code Division Multiple Access, CDMA), wideband code division multiple access (Wideband Code Division Multiple Access, WCDMA), long term evolution (Long Term Evolution, LTE), email, short message service (Short Messaging Service, SMS), and the like.

The memory 1120 may be used to store software programs and modules, and the processor 1180 executes the software programs and modules stored in the memory 1120 to perform various functional applications and data processing of the cellular phone. The memory 1120 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, phonebook, etc.) created according to the use of the handset, etc. In addition, memory 1120 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device.

The input unit 1130 may be used to receive input numerical or character information and generate key signal inputs related to object settings and function control of the mobile phone. In particular, the input unit 1130 may include a touch panel 1131 and other input devices 1132. The touch panel 1131, also referred to as a touch screen, may collect touch operations on or near an object (e.g., the object's operation on the touch panel 1131 using any suitable object or accessory such as a finger, a stylus, etc., or near the touch panel 1131) and drive the corresponding connection device according to a predetermined program. Alternatively, the touch panel 1131 may include two parts, a touch detection device and a touch controller. The touch detection device detects the touch azimuth of the object, detects a signal brought by touch operation and transmits the signal to the touch controller; the touch controller receives touch information from the touch sensing device and converts it into touch point coordinates, which are then sent to the processor 1180, and can receive commands from the processor 1180 and execute them. In addition, the touch panel 1131 may be implemented in various types such as resistive, capacitive, infrared, and surface acoustic wave. The input unit 1130 may include other input devices 1132 in addition to the touch panel 1131. In particular, other input devices 1132 may include, but are not limited to, one or more of a physical keyboard, function keys (e.g., volume control keys, switch keys, etc.), a trackball, mouse, joystick, etc.

The display unit 1140 may be used to display information input by an object or information provided to the object and various menus of a mobile phone. The display unit 1140 may include a display panel 1141, and optionally, the display panel 1141 may be configured in the form of a liquid crystal display (Liquid Crystal Display, LCD), an Organic Light-Emitting Diode (OLED), or the like. Further, the touch panel 1131 may overlay the display panel 1141, and when the touch panel 1131 detects a touch operation thereon or thereabout, the touch panel is transferred to the processor 1180 to determine the type of touch event, and then the processor 1180 provides a corresponding visual output on the display panel 1141 according to the type of touch event. Although in fig. 11, the touch panel 1131 and the display panel 1141 are two separate components to implement the input and output functions of the mobile phone, in some embodiments, the touch panel 1131 may be integrated with the display panel 1141 to implement the input and output functions of the mobile phone.

The handset may also include at least one sensor 1150, such as a light sensor, motion sensor, and other sensors. Specifically, the light sensor may include an ambient light sensor and a proximity sensor, wherein the ambient light sensor may adjust the brightness of the display panel 1141 according to the brightness of ambient light, and the proximity sensor may turn off the display panel 1141 and/or the backlight when the mobile phone moves to the ear. As one of the motion sensors, the accelerometer sensor can detect the acceleration in all directions (generally three axes), and can detect the gravity and direction when stationary, and can be used for applications of recognizing the gesture of a mobile phone (such as horizontal and vertical screen switching, related games, magnetometer gesture calibration), vibration recognition related functions (such as pedometer and knocking), and the like; as for other sensors such as gyroscopes, barometers, hygrometers, thermometers, infrared sensors, etc. that may be further configured in the mobile phone, details are not described here.

Audio circuitry 1160, speaker 1161, and microphone 1162 may provide an audio interface between the object and the handset. The audio circuit 1160 may transmit the received electrical signal converted from audio data to the speaker 1161, and may be converted into a sound signal by the speaker 1161 to be output; on the other hand, the microphone 1162 converts the collected sound signals into electrical signals, which are received by the audio circuit 1160 and converted into audio data, which are processed by the audio data output processor 1180 for transmission to, for example, another cell phone via the RF circuit 1110, or which are output to the memory 1120 for further processing.

WiFi belongs to a short-distance wireless transmission technology, and a mobile phone can help an object to send and receive emails, browse webpages, access streaming media and the like through a WiFi module 1170, so that wireless broadband Internet access is provided for the object. Although fig. 11 shows a WiFi module 1170, it is understood that it does not belong to the necessary constitution of the handset.

The processor 1180 is a control center of the mobile phone, and connects various parts of the entire mobile phone using various interfaces and lines, and performs various functions and processes of the mobile phone by running or executing software programs and/or modules stored in the memory 1120 and calling data stored in the memory 1120, thereby performing overall monitoring of the mobile phone. In the alternative, processor 1180 may include one or more processing units; preferably, the processor 1180 may integrate an application processor and a modem processor, wherein the application processor primarily processes an operating system, an object interface, an application program, etc., and the modem processor primarily processes wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 1180.

The handset further includes a power supply 1190 (e.g., a battery) for powering the various components, which may be logically connected to the processor 1180 via a power management system so as to provide for the management of charging, discharging, and power consumption by the power management system.

Although not shown, the mobile phone may further include a camera, a bluetooth module, etc., which will not be described herein.

In the embodiment of the present application, the processor 1180 included in the terminal is configured to perform the embodiment shown in fig. 3 and the respective embodiments corresponding to fig. 3.

There is also provided in an embodiment of the present application a computer readable storage medium having stored therein a computer program which, when run on a computer, causes the computer to perform the steps performed by the server in the method described in the embodiment of fig. 3 as described above.

There is also provided in an embodiment of the present application a computer program product comprising a program which, when run on a computer, causes the computer to perform the steps performed by the server in the method described in the embodiment of fig. 3 as described above.

It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.

In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the elements is merely a logical functional division, and there may be additional divisions when actually implemented, e.g., at least two elements or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on at least two network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a read-only memory (ROM), a random access memory (random access memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

The above embodiments are merely for illustrating the technical solution of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the corresponding technical solutions.

Claims

1. A method of data processing, comprising:

acquiring first object data of an object to be predicted in a first time period, wherein the first object data comprises first object tolerance and first object information, the first object tolerance is used for representing tolerance degree of the object to be predicted for reducing corresponding resource quantity of a product owned by the object to be predicted in the first time period, and the first object information is related behavior information of the object to be predicted for the product in the first time period;

acquiring a first predicted data tolerance of the object to be predicted through a data tolerance prediction model based on a first data characteristic of the first object data;

and acquiring the first predicted object tolerance of the object to be predicted in a second time period based on the first predicted data tolerance of the object to be predicted and the first object tolerance through an object tolerance prediction model, wherein the second time period is the next time period adjacent to the first time period, and the first predicted object tolerance is used for representing the predicted tolerance degree of the object to be predicted, which is reduced in the amount of resources corresponding to products owned by the object to be predicted in the second time period.

2. The method according to claim 1, wherein the method further comprises:

acquiring market information in the first time period;

determining a market prediction influence degree of the object to be predicted based on the market information in the first time period and the first object data, wherein the market prediction influence degree indicates influence of a trading market on the behavior of the object to be predicted in the first time period;

the obtaining, by an object tolerance prediction model, the first predicted object tolerance of the object to be predicted in a second time period based on the first predicted data tolerance of the object to be predicted and the first object tolerance includes:

and acquiring the first predicted object tolerance of the object to be predicted in the second time period through the object tolerance prediction model based on the first predicted data tolerance of the object to be predicted, the first object tolerance and the market prediction influence degree of the object to be predicted.

3. The method of claim 2, wherein the first object data further comprises object base information;

the characterizing the first object data to determine a first data feature of the first object data includes:

Carrying out characterization processing on the first object data, and determining first data characteristics of the object to be predicted and object image characteristics of the object to be predicted;

the obtaining, based on the data characteristics of the first object data, the first predicted data tolerance of the object to be predicted through a data tolerance prediction model includes:

and acquiring the first prediction data tolerance of the object to be predicted through the data tolerance prediction model based on the first data characteristic of the first object data and the object portrait characteristic of the object to be predicted.

4. The method according to claim 1, wherein the method further comprises:

acquiring second object data of the object to be predicted in the second time period, wherein the second object data comprises second object tolerance and second object information, the second object tolerance is used for representing the tolerance degree of the object to be predicted to the reduction of the corresponding resource amount of a product owned by the object to be predicted in the second time period, and the second object information is related behavior information of the object to be predicted to the product in the second time period;

Characterizing the second object data to determine second data characteristics of the object to be predicted;

acquiring a second predicted data tolerance of the object to be predicted through the data tolerance prediction model based on a second data characteristic of the second object data;

and acquiring a second predicted object tolerance of the object to be predicted in a third time period based on a second predicted data tolerance of the second object data and the second object tolerance through the object tolerance prediction model, wherein the third time period is a next time period adjacent to the second time period, and the second predicted object tolerance is used for representing a predicted tolerance degree of the object to be predicted for reducing a resource amount corresponding to a product owned by the object to be predicted in the third time period.

5. The method of claim 4, wherein the first predicted object tolerance is the second object tolerance.

6. The method of claim 4, wherein the first object tolerance is determined based on a questionnaire filled in by the object to be predicted during the first time period;

The method further comprises the steps of:

7. The method of claim 4, wherein the second object information comprises second object transaction information or the second object information comprises the second object transaction information and the second object access information;

the method further comprises the steps of:

determining a first data tolerance of the object to be predicted within the second period by a redemption policy based on the second object transaction information in the second object information;

the first data tolerance is determined as the second object tolerance.

8. The method of claim 4, wherein the second object information comprises second object binning information or the second object information comprises the second object binning information and the second object access information;

The method further comprises the steps of:

determining a second data tolerance of the object to be predicted in the second period based on the second object holding information in the second object information;

9. The method of claim 1, wherein the data features include at least one of an access feature, a purchase feature, and a holding feature, the access feature being derived based on object access information, the purchase feature being derived based on object transaction information, and the holding feature being derived based on object holding information;

the access data tolerance prediction model is used for obtaining the access prediction data tolerance of the object based on the access characteristic;

the purchasing data tolerance prediction model is used for obtaining purchasing prediction data tolerance of the object based on the purchasing characteristics;

The first prediction data tolerance of the object to be predicted comprises at least one of access prediction data tolerance of the object to be predicted, buying prediction data tolerance of the object to be predicted and holding prediction data tolerance of the object to be predicted.

10. The method according to claim 1, wherein the method further comprises:

screening a target data sample set from the initial data sample set based on a redemption policy and a holding policy, wherein the target data sample set includes target data samples of a plurality of object samples;

determining a target data tolerance for each subject sample based on the target set of data samples by the redemption policy and the binning policy;

And training the data tolerance prediction model to be trained based on the target data tolerance of each object sample and the predicted data tolerance of each object sample to obtain the data tolerance prediction model.

11. The method according to claim 10, wherein the method further comprises:

and training the object tolerance prediction model to be trained based on the object tolerance of each object sample and the predicted object tolerance of each object sample to obtain the object tolerance prediction model.

12. The method of claim 11, wherein the method further comprises:

determining a market prediction influence degree of each object sample in different periods based on market information in each different period and target data samples of the plurality of object samples, wherein the market prediction influence degree indicates influence of a market on behaviors of the object samples in one time period;

The obtaining the predicted object tolerance of each object sample through the object tolerance prediction model to be trained based on the predicted data tolerance of each object sample and the object tolerance of each object sample comprises the following steps:

13. An object tolerance prediction apparatus, comprising:

the device comprises an acquisition module, a prediction module and a prediction module, wherein the acquisition module is used for acquiring first object data of an object to be predicted in a first time period, and the first object data comprises first object tolerance and first object information;

the obtaining module is further configured to obtain a first predicted data tolerance of the object to be predicted through a data tolerance prediction model based on a first data feature of the first object data;

The obtaining module is further configured to obtain, based on the first predicted data tolerance of the object to be predicted and the first object tolerance, a first predicted object tolerance of the object to be predicted in a second time period through an object tolerance prediction model, where the second time period is a next time period adjacent to the first time period.

14. A computer device, comprising: memory, transceiver, processor, and bus system;

wherein the memory is used for storing programs;

the processor being adapted to execute a program in the memory to implement the method of any one of claims 1 to 12;

the bus system is used for connecting the memory and the processor so as to enable the memory and the processor to communicate.

15. A computer readable storage medium comprising instructions which, when run on a computer, cause the computer to perform the method of any one of claims 1 to 12.

16. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the method of any of claims 1 to 12.