CN114077962A

CN114077962A - Resource quota re-determination method and device based on user tag and computer equipment

Info

Publication number: CN114077962A
Application number: CN202111213765.XA
Authority: CN
Inventors: 程锋; 丁楠; 苏绥绥
Original assignee: Beijing Qilu Information Technology Co Ltd
Current assignee: Beijing Qilu Information Technology Co Ltd
Priority date: 2021-10-19
Filing date: 2021-10-19
Publication date: 2022-02-22

Abstract

The invention provides a resource quota re-determining method and device based on a user tag and computer equipment. The method comprises the following steps: determining a reference user and user characteristics thereof, and performing cluster calculation on the users in the candidate user set and the reference user to screen out a target user group; regenerating a user label based on the user characteristics of the screened target user group, and establishing a training data set for training a resource quota determination model according to the user label; inputting the user characteristics of the candidate users into the trained resource quota determining model, and outputting the resource quota corresponding to the candidate users. According to the method and the device, the user label is regenerated to define the good and bad samples to establish the training data set, and the training data set is used for training the resource quota determining model, so that the change condition of the demand of the user on the internet resources can be effectively quantified, the resource quota of the candidate user can be determined more accurately, and the distribution process of the internet resources can be further optimized.

Description

Resource quota re-determination method and device based on user tag and computer equipment

Technical Field

The invention relates to the field of computer information processing, in particular to a resource quota re-determining method and device based on a user tag and computer equipment.

Background

In internet-based application technology, there is often a need to exchange resources between different parties. Resources, as referred to herein, refer to any available material, information, money, time, etc. Information resources include computing resources and various types of data resources. The data resources include various private data in various domains. In the process of allocating resources, it is often necessary to authenticate a resource allocation right of a user and allocate different resource quotas to different users, where the resource allocation right refers to authentication of whether the user has the right to acquire resources, and may be authenticated by a specific resource management mechanism or by all parties of the resources. By resource quota is meant the highest amount of resources that the user can obtain within a particular time.

Money-related resources are also commonly referred to as internet resources, which refers to the sum or aggregate of a series of objects related to the structure, quantity, scale, distribution, effect and interaction relationship of resource service subjects and objects in the internet field, and only when the internet resources are configured efficiently in production and life, the internet resources and economic sustainable development can be realized. For companies that provide internet resource services, the internet resource may be the total amount of funds, or the amount of assets equivalent to funds, and so on. For companies that provide internet resource services, some of the internet assets can be used to provide resource services to individual users, some can be used to provide financial services to other enterprise users, and others can be used to invest in the development of the company or perform other internet related business.

For companies of internet resource services, it is important how to reasonably allocate internet resources among different businesses, since the total internet resources are limited in a relatively fixed time. For enterprise users or other financial related businesses served by an internet resource service company, the time and period for occupying internet resources can be generally approved through planning and approval in advance, which is favorable for overall arrangement of allocation of internet resources. For individual users, because of individual differences of individual users, the internet resource service company can hardly predict the plan and time of internet resource demand of individual users in advance, and the total amount of available or usable internet resources changes in different time periods, so how to effectively quantify the demand change of individual users for internet resource services over time, make more reasonable allocation of internet resources for individual users, and effectively improve the recyclability and reusability of internet resources is a difficult problem faced by current internet resource service companies.

Therefore, there is a need to provide a more optimized resource quota re-determination method.

Disclosure of Invention

The method aims to solve the technical problems that the demand change of each user on the internet resource service along with the time change cannot be accurately quantified, the distribution process of the internet resources is optimized, and the recoverability and the reusability of the internet resources are effectively improved in the prior art.

A first aspect of the present invention provides a method for re-determining a resource quota based on a user tag, including: determining a reference user and user characteristics thereof, and performing cluster calculation on the users in the candidate user set and the reference user to screen out a target user group; regenerating a user label based on the user characteristics of the screened target user group, and establishing a training data set for training a resource quota determination model according to the user label; inputting the user characteristics of the candidate users into the trained resource quota determining model, and outputting the resource quota corresponding to the candidate users.

According to an alternative embodiment, said determining the reference user and its user characteristics comprises: determining a reference user with a resource quota value larger than that of the candidate user in the candidate user set; and determining the user basic information characteristics, the professional information characteristics and the resource return performance characteristics of the reference user.

According to an optional embodiment, the clustering the users in the candidate user set with the reference user to filter out the target user group includes: generating a plurality of user groups after the clustering calculation, and screening a target user group from the plurality of user groups according to a screening rule; wherein the screening rule includes that the total number of users in the user group is greater than a certain number, the ratio of the number of the reference users to the number of the candidate users in the user group is within a predetermined proportion range, and the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is compared.

According to an optional embodiment, when the total number of users of the user group is greater than a certain number and the ratio of the number of the reference users to the number of the candidate users in the user group is 1:1, the user group is determined to be a target user group.

According to an optional embodiment, when the total number of users in the user group is greater than a certain number, and the ratio of the number of the reference users to the number of the candidate users in the user group is 1:1, and/or the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is greater than a set value, the user group is determined to be a target user group.

According to an optional embodiment, the generating a user tag based on the user characteristics of the screened target user group includes: redefining user labels which are similar to the candidate users in the target user group and become 1, and redefining user labels which are dissimilar to the candidate users in the target user group and become 0 to establish a training data set; and/or redefining the user tags with the similarity greater than a specified value to the candidate users in the target user group to be 1, and redefining the positive samples and the negative samples to the user tags with the similarity less than or equal to the specified value to be 0, so as to establish a training data set.

According to an optional implementation mode, a K-means algorithm is used for carrying out clustering calculation or multi-round clustering calculation on the users in the candidate user set and the reference user.

According to an optional embodiment, constructing the built resource quota determining model comprises: establishing a resource quota determining model by using a logistic regression model, an Xgboost model and/or a deep neural network, wherein the resource quota determining model calculates a user evaluation value, and outputs a resource quota value or a resource quota range corresponding to the user evaluation value of the candidate user, and the user evaluation value is a numerical value between 0 and 1; or the resource quota determining model directly outputs the resource quota value or the resource quota range of the candidate user.

In addition, a second aspect of the present invention further provides a resource quota re-determination apparatus based on a user tag, including: the screening module is used for determining reference users and user characteristics thereof, and performing clustering calculation on the users in the candidate user set and the reference users so as to screen out a target user group; the establishing module is used for regenerating a user label based on the user characteristics of the screened target user group and establishing a training data set for training a resource quota determining model according to the user label; and the calculation determining module is used for inputting the user characteristics of the candidate users into the trained resource quota determining model and outputting the resource quotas corresponding to the candidate users.

Furthermore, a third aspect of the present invention also provides a computer device, comprising a processor and a memory, wherein the memory is used for storing a computer executable program, and when the computer program is executed by the processor, the processor executes the resource quota re-determination method based on the user tag according to the present invention.

Furthermore, a fourth aspect of the present invention further provides a computer program product, which stores a computer executable program, and when the computer executable program is executed, the method for re-determining a resource quota based on a user tag according to the present invention is implemented.

Advantageous effects

Compared with the prior art, the method and the device have the advantages that the reference user and the user characteristics thereof are determined, the users in the candidate user set and the reference user are subjected to clustering calculation, the target user group is screened out, and the target user group can be accurately screened out from the obtained multiple user groups; the user label definition good or bad sample regenerated by using the user characteristics of the screened target user group is used for establishing a training data set, and the resource quota determining model is trained by using the training data set, so that the requirement change condition of the user for the internet resource can be effectively quantified, and the resource quota can be accurately determined.

Furthermore, the target user group is screened according to the screening rule, so that the target user group can be screened from the obtained multiple user groups more accurately; by calculating and determining the adjustment coefficient of the resource quota, and according to the adjustment coefficient and the determined resource quota interval, the resource quota of the candidate user can be re-determined according to the requirement change of each user on the internet resource service along with the time change, the resource quota can be more accurately determined, the user resource quota determining method can be further optimized, the distribution process of the internet resources can be further optimized, and the recyclability and the reusability of the internet resources can be effectively improved.

Drawings

In order to make the technical problems solved by the present invention, the technical means adopted and the technical effects obtained more clear, the following will describe in detail the embodiments of the present invention with reference to the accompanying drawings. It should be noted, however, that the drawings described below are only illustrations of exemplary embodiments of the invention, from which other embodiments can be derived by those skilled in the art without inventive faculty.

Fig. 1 is a flowchart of an example of a resource quota re-determination method based on a user tag according to the present invention.

FIG. 2 is a flowchart of another example of a user tag-based resource quota re-determination method of the present invention.

FIG. 3 is a flowchart of another example of a user tag-based resource quota re-determination method of the present invention.

Fig. 4 is a schematic block diagram of an example of the apparatus for re-determining a resource quota based on a user tag according to the present invention.

Fig. 5 is a schematic block diagram illustrating another example of the apparatus for re-determining a resource quota based on a user tag according to the present invention.

Fig. 6 is a schematic block diagram illustrating still another example of the apparatus for re-determining a resource quota based on a user tag according to the present invention.

Fig. 7 is a block diagram of an exemplary embodiment of a computer device according to the present invention.

Fig. 8 is a block diagram of an exemplary embodiment of a computer program product according to the present invention.

Detailed Description

Exemplary embodiments of the present invention will now be described more fully with reference to the accompanying drawings. The exemplary embodiments, however, may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these exemplary embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the invention to those skilled in the art. The same reference numerals denote the same or similar elements, components, or parts in the drawings, and thus their repetitive description will be omitted.

Features, structures, characteristics or other details described in a particular embodiment do not preclude the fact that the features, structures, characteristics or other details may be combined in a suitable manner in one or more other embodiments in accordance with the technical idea of the invention.

In describing particular embodiments, the present invention has been described with reference to features, structures, characteristics or other details that are within the purview of one skilled in the art to provide a thorough understanding of the embodiments. One skilled in the relevant art will recognize, however, that the invention may be practiced without one or more of the specific features, structures, characteristics, or other details.

The flow charts shown in the drawings are merely illustrative and do not necessarily include all of the contents and operations/steps, nor do they necessarily have to be performed in the order described. For example, some operations/steps may be decomposed, and some operations/steps may be combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.

The block diagrams shown in the figures are functional entities only and do not necessarily correspond to physically separate entities. I.e. these functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor means and/or microcontroller means.

It will be understood that, although the terms first, second, third, etc. may be used herein to describe various elements, components, or sections, these terms should not be construed as limiting. These phrases are used to distinguish one from another. For example, a first device may also be referred to as a second device without departing from the spirit of the present invention.

The term "and/or" and/or "includes any and all combinations of one or more of the associated listed items.

In view of the above problems, the present invention provides a resource quota re-determination method based on a user tag, which performs cluster calculation on users in a candidate user set and a reference user by determining the reference user and user characteristics thereof, screens out a target user group, and can screen out the target user group from a plurality of obtained user groups more accurately; the user label regenerated by using the user characteristics of the screened target user group defines a good-bad sample to establish a training data set, and the resource quota determining model is trained by using the training data set, so that the requirement change condition of the user on internet resources can be effectively quantified, the model parameters of the resource quota determining model can be optimized, and the model accuracy can be improved.

In order that the objects, technical solutions and advantages of the present invention will become more apparent, the present invention will be further described in detail with reference to the accompanying drawings in conjunction with the following specific embodiments. For convenience, the implementation of the resource quota re-determination method is illustrated in the present invention by taking internet data resources as an example, but those skilled in the art should understand that the present invention can also be used for re-determination of other resource quotas.

Example 1

Hereinafter, an embodiment of the resource quota re-determination method based on the user tag of the present invention will be described with reference to fig. 1 to 3.

As shown in fig. 1, the resource quota re-determination method includes the following steps.

Step S101, determining reference users and user characteristics thereof, and performing clustering calculation on the users in the candidate user set and the reference users to screen out a target user group.

And step S102, regenerating a user label based on the user characteristics of the screened target user group, and establishing a training data set for training a resource quota determination model according to the user label.

Step S103, inputting the user characteristics of the candidate users into the trained resource quota determination model, and outputting the resource quota corresponding to the candidate users.

It should be noted that the resource quota re-determination method of the present invention is particularly suitable for an application scenario of an internet resource service. The internet resource service comprises the service resources such as shopping, riding, maps, taking out, sharing single vehicles and the like which are provided by the application of user equipment or a client to the internet resource service platform. For example, the internet resource service includes a resource usage service, a resource allocation service, a resource raising service, a resource guarantee service or a mutual aid service, a group buying and taking service, and the like. Where resources refer to any available substances, information, time, information resources including computing resources and various types of data resources. The data resources include various private data in various domains.

First, in step S101, a reference user and user characteristics thereof are determined, and clustering calculation is performed between users in a candidate user set and the reference user to filter out a target user group.

Specifically, according to the business targets of different internet resource services, the reference user is determined from the pre-classified user group.

More specifically, the business objectives include the total amount of internet resources, the amount of resources that can flow, adjusting (increasing or decreasing) the resource quotas, etc. over a particular time. The pre-classified user groups comprise user groups divided according to parameters such as city, age, gender, income and/or motor activity.

In one embodiment, in an application scenario of the internet resource allocation service, when a business goal is to adjust (increase or decrease) a resource quota, a reference user and a user characteristic thereof are determined from a pre-classified user group of the internet resource allocation service.

Specifically, the reference subscriber includes a subscriber whose resource allocation quota is above a specific value or within a specific range. For example, the resource allocation limit is, for example, 1 ten thousand yuan or more or in the range of 1 ten thousand yuan to 2 ten thousand yuan.

More specifically, for example, in a case that a service target is to increase the resource quota of a user, a reference user larger than the resource quota value of a candidate user in the candidate user set is determined; and determining the user basic information characteristics, the professional information characteristics and the resource return performance characteristics of the reference user.

Further, the user characteristics include a user account, a mobile phone number, a user equipment ID, user basic information, performance data of internet resource service usage, and the like. The basic information of the user comprises information data of the gender, the age, the income, the city and the like of the user.

Specifically, the performance data of the internet resource service usage includes the internet resource service type selected by the user, the resource quota corresponding to each internet resource service type, the resource usable time, whether the resource return is overdue or violated, and the like.

It should be noted that the above description is only given by way of example, and the present invention is not limited thereto. In other embodiments, the resource quota data further comprises resource return time, number of times the resource has been returned, amount of unreturned resources, overdue data for resource returns, default data, fraud data, number of times the resource quota was adjusted within a particular time, and the like. In addition, the number of resource quota uses, the number of resource quota adjustments, the data related to the shared resource credential (e.g., whether there is a shared resource credential, the number of times the shared resource credential is received, or the number of times the shared resource credential is received within a certain time, a change characteristic from not having a shared resource credential to having a shared resource credential within a certain time), and the like may also be included.

Further, according to the determined reference user and the user characteristics thereof, clustering calculation is carried out on the users in the candidate user set and the reference user or the clustering calculation is carried out on the users in the candidate user set and the reference user.

It should be noted that, in the present invention, the candidate user set refers to a user group or a user set that has used internet resources provided by an internet resource service.

In one embodiment, for example, the determined resource quota for the reference user is 1.6 ten thousand dollars, and the user characteristics include user age, academic history, and city. For example, screening users from a set of candidate users having a resource quota of less than 1.6 ten thousand dollars.

In another embodiment, for example, the determined resource quota of the reference user is 1.8 ten thousand yuan, and the user characteristics include user gender, academic history, and performance data of internet resource service usage. For example, screening users from a candidate user set with a resource quota of 1.8 ten thousand dollars or less.

In yet another embodiment, for example, the determined resource quota for the reference user is 2 ten thousand dollars, and the user characteristics include user learned history and resource quota data. For example, screening users from a candidate user set with a resource quota of 2 ten thousand dollars or less.

In yet another embodiment, for example, the determined resource quota for the reference user is 1 ten thousand dollars, and the user characteristics include user learned history, resource return overdue data, and breach data. For example, screening users from a set of candidate users with a resource quota of less than 1 ten thousand dollars.

It should be noted that the above description is only given by way of example, and the present invention is not limited thereto.

Specifically, according to the user characteristics of the reference user, a K-means algorithm is used for carrying out clustering calculation or multi-round clustering calculation on the users in the candidate user set and the reference user. For example, two or more rounds of clustering calculation are performed, and a plurality of user groups are generated after the clustering calculation so as to screen out a target user group from the plurality of user groups.

For multi-round clustering calculation, each round of clustering adopts different K values, and an initial clustering result is output, wherein the K value is the number of initial central vectors in a K-means algorithm.

As a specific embodiment, the user samples are set to obey a normal distribution, the number of groups (or the number of classes) in the initial clustering result increases as the number of rounds of the multi-round clustering process increases, and the number of classes in the initial clustering result gradually decreases when the number of rounds of the multi-round clustering process reaches a specific number of rounds.

Specifically, the clustering calculation is performed using an algorithm of a centroid, for example. For the Toronto clustering process, different k values are adopted for each round of clustering, and an initial clustering result is output. Wherein, each round of clustering processing comprises the following steps: setting an initial k value; randomly generating K class center vectors, and iteratively updating the class center vectors by using a K-means algorithm until the distance between the class center vector in the current iteration and the class center vector in the last iteration is less than a specified threshold value.

More specifically, for example, the euclidean distance of each user sample to each class center vector is calculated; in Euclidean distances from each user sample to various central vectors, taking the class of the class central vector with the minimum distance as the class to which the sample belongs in the current iteration; and taking the mean vector of the samples belonging to the same class as the class center vector of the next iteration. Thus, after multiple rounds of clustering calculation, multiple user groups are obtained.

It should be noted that the clustering algorithm is not limited to this, and the above description is only an example, and is not to be construed as limiting the present invention, and in other examples, an algorithm based on probability distribution, an EM algorithm, or the like may be used.

And further, screening a target user group from the obtained plurality of user groups according to a screening rule.

Specifically, the screening rule includes that the total number of users in the user group is greater than a certain number, the ratio of the number of the reference users to the number of the candidate users in the user group is within a predetermined ratio, and the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is compared.

The similarity algorithm is, for example, to calculate the similarity of user vectors generated by user features, or calculate euclidean distances of the user vectors, but the algorithm is not limited to this, and the above description is only an example, and is not to be construed as limiting the present invention.

In one embodiment, for example, the filtering rule is that the total number of users in the user group is greater than a certain number, and the ratio of the number of the reference users to the number of the candidate users in the user group is, for example, 1: 1.

In another embodiment, for example, the screening rule is that the total number of users in the user group is greater than a certain number, and the ratio of the number of the reference users to the number of the candidate users in the user group is, for example, in a predetermined ratio range greater than 3:4 to 9: 10.

In yet another embodiment, for example, the screening rule is that the total number of users in the user group is greater than a certain number, the ratio of the number of the reference users to the number of the candidate users in the user group is, for example, 1:1, and the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is compared.

For the determination of the target user group, in an embodiment, when the total number of users in the user group is greater than a specific number and the ratio of the number of the reference users to the number of the candidate users in the user group is 1:1, the user group is determined to be the target user group.

In another embodiment, when the total number of users in the user group is greater than a certain number, and the ratio of the number of the reference users to the number of the candidate users in the user group is 1:1, and/or the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is greater than a set value, the user group is determined to be a target user group.

Therefore, according to the screening rule, the target user group can be screened from the obtained multiple user groups more accurately.

Next, in step S102, a user label is regenerated based on the user characteristics of the screened target user group, and a training data set for training the resource quota determination model is established according to the user label.

Specifically, the user tag is regenerated according to the user characteristics of the screened target user group, where the screened users are users who need to adjust (increase or decrease) the resource quota.

For example, the user characteristics of the filtered target user group include the city of the user, performance data of the internet resource service usage, and the like. The performance data of the internet resource service usage comprises the internet resource service type selected by the user, the resource quota corresponding to each internet resource service type, the resource usable time, the resource quota using times, the resource quota adjusting times, the data related to the special resource voucher (for example, whether the special resource voucher exists, the getting times of the special resource voucher or the getting times of the special resource voucher in a specific time, the change characteristic from not holding the special resource voucher to holding the special resource voucher in the specific time), whether the resource return is overdue or violated, and the like.

It should be noted that, in the present invention, the special resource usage certificate includes a card, a coupon and/or a discount coupon of the resource amount that can be used in a specified time. For example, the specified time includes 1 month, 2 months, 3 months, 6 months, 12 months, and the like. For example, the usable resource quota is 1000 yuan, 2000 yuan, 5000 yuan, or the like, or a quota use range of 1000 yuan to 5000 yuan, 1000 yuan to 3000 yuan, 2000 yuan to 10000 yuan, or the like.

In one embodiment, the user label of 1 is re-generated for users similar to the candidate user in the target user group, and the positive and negative examples are re-defined for users dissimilar to the candidate user in the target user group to be the user label of 0, so as to establish the training data set.

In another embodiment, the user tags with similarity greater than a specified value (e.g., 70%, 80%, or 90% similarity) to the candidate users in the target user group are re-generated as 1, and the user tags with similarity less than or equal to the specified value (e.g., 70%, 80%, or 90% similarity) to the candidate users in the target user group are re-defined as positive and negative examples to establish the training data set.

In another embodiment, the training data set is created by redefining the user tags whose similarity to the candidate users in the target user group is greater than a predetermined value (e.g., the predetermined value is an arbitrary value of 70% to 90% similarity) and whose users have the shard resource voucher (or whose change from not holding to holding the shard resource usage voucher occurs within a predetermined time period) are reset to 1, and by redefining the user tags whose similarity to the candidate users in the target user group is less than or equal to the predetermined value (e.g., the predetermined value is an arbitrary value of 70% to 90% similarity) and whose users have no shard resource voucher are reset to 0. Wherein, 1 indicates that the user has a demand change that needs to adjust the resource quota, and 0 indicates that the user has no demand change that needs to adjust the resource quota. Generally, the closer the user tag is to 1, the greater the probability that the user needs to adjust the resource quota, and the closer the user tag is to 0, the less the probability that the user needs to adjust the resource quota.

It should be noted that the above is only an example, and the invention is not limited thereto, wherein the specific time period includes 6 months, 8 months, 10 months, 12 months, 18 months, etc. In addition, in other embodiments, the probability of the demand change may also be determined using a number of parameters such as the number of times a coupon, a discount coupon, or the like is issued for a particular resource usage credential and a coupon value within a particular time period. The foregoing is described as an alternative example and is not to be construed as limiting the present invention.

Specifically, the special resource use voucher comprises a card, a coupon and/or a discount coupon of the resource limit which can be used in a specified time. For example, the specified time includes 1 month, 2 months, 3 months, 6 months, 12 months, and the like.

As a specific implementation manner, the training data set includes historical users labeled with the generated user labels and user feature data thereof, where the user features include user account numbers, mobile phone numbers, user equipment IDs, user basic information, and the like.

Specifically, the user basic information includes information data such as user gender, age, income and the like.

In another embodiment, the training data set includes historical users tagged with the generated user tags and their user characteristic data, as well as resource quota data.

Specifically, the resource quota data includes internet resource service types selected by the user, resource quotas corresponding to the internet resource service types, resource usable time, resource quota using times, resource quota adjusting times, data related to the shared resource voucher (for example, whether the shared resource voucher exists, the number of times of getting the shared resource voucher or the number of times of getting the shared resource voucher in a specific time, a change characteristic from not holding the shared resource voucher to holding the shared resource voucher in the specific time), whether resource return is overdue or violated, and the like.

In yet another embodiment, the resource quota data includes internet resource service types selected by the user, resource quotas corresponding to the respective internet resource service types, resource usable time, resource return time, number of times the resources have been returned, amount of unreturned resources, overdue data of resource returns, default data, fraud data, and the like.

It should be noted that the foregoing is only an example, and is not to be construed as limiting the present invention, and in other embodiments, the resource quota data may further include whether multiple internet resource service types are included, the number of service types, whether fraud data exists, and the like.

Optionally, the resource quota determining model includes: the resource quota determination model is built, for example, using a logistic regression model, an Xgboost model, and/or a deep neural network. And the resource quota determining model calculates a user evaluation value, and outputs a resource quota value or a resource quota range corresponding to the user evaluation value of the candidate user, wherein the user evaluation value is a numerical value between 0 and 1.

It should be noted that the above description is only given by way of example, and the present invention is not limited thereto. In other embodiments, the resource quota determination model may also directly output a resource quota value or a resource quota range for the candidate user.

Further, the established resource quota determination model is trained using a training data set.

Optionally, for the model training process, optimizing model parameters including weight parameters and bias parameters between layers of the deep neural network is further included.

It should be noted that the above description is only given by way of example, and the present invention is not limited thereto. In other examples, the model parameters also include the number of layers of the deep neural network, the number of iterations, and the learning rate.

Specifically, for example, the prior probability of the model parameter is set to comply with the normal distribution, the MCMC method is used to sample the distribution of the weight parameter and the bias parameter for multiple times to obtain a parameter combination set, and more specifically, when the parameter combination set is used for testing, the parameter combination set is used to perform multiple forward propagation on the input same user characteristic data, calculate the mean and the statistical variance of each model parameter to optimize each model parameter, and finally output the neural network after parameter optimization. Therefore, the model parameters of the resource quota determination model can be optimized, and the model accuracy can be improved.

Therefore, the user label definition good or bad sample regenerated by using the user characteristics of the screened target user group is used for establishing the training data set, and the resource quota determining model is trained by using the training data set, so that the requirement change condition of the user on the internet resource can be effectively quantified, the model parameters of the resource quota determining model can be optimized, and the model accuracy can be improved.

Next, in step S103, the user characteristics of the candidate user are input into the trained resource quota determination model, and the resource quota corresponding to the candidate user is output.

Specifically, user characteristics of a candidate user are obtained, and the user characteristics of the candidate user are input into a trained resource quota determination model to output a resource quota corresponding to the candidate user.

As shown in fig. 2, in another embodiment, the resource quota re-determining method further includes a step S201 of determining a target user group.

In step S201, a target user group (i.e., a candidate user group) is determined, and candidate users to be adjusted internet resource quota (also referred to as resource quota or resource allocation amount in the present invention) are screened from the target user group.

In step S201, a coefficient determination curve is fitted for determining the first calculation coefficient a.

In the example shown in fig. 2, step S103 in fig. 1 is divided into step S103 and step S201. Since step S101, step S102, and step S103 in fig. 2 are substantially the same as step S101, step S102, and step S103 in fig. 1, descriptions of the same parts are omitted.

Specifically, under the scene of screening internet resource allocation services (such as financial services) for improving resource quotas, candidate user groups within a specified resource quota range are determined. For example, the specified resource quota range is, for example, 2000 yuan to 5000 yuan, 5000 yuan to 8000 yuan, or the like.

Further, a step of screening candidate users from the candidate user set (i.e. the target user group) is also included.

Specifically, user characteristic data, resource quota related data (for example, the number of times of issuing the special resource credential or the number of times of using the special resource credential within a specific time), an overdue probability or a resource return probability of historical users of different internet resource services are obtained, and a first user score of each historical user is calculated.

In one embodiment, a curve is determined according to a first user score of the historical user and a fitting coefficient of the number of times of issuance of the special resource voucher (or the number of times of usage of the special resource voucher) in a specific time, and the coefficient determination curve is used for determining a first calculation coefficient a of the resource quota increase amplitude (or decrease amplitude) of the user.

It should be noted that, in the above example, the number of times of issuance of the special resource voucher (or the number of times of usage of the special resource voucher) within a specific time is used to characterize the usage rate of the special resource voucher, wherein the specific time includes 6 months, 9 months, 12 months, 18 months or 24 months. However, the present invention is not limited thereto, and the above description is only by way of example and is not to be construed as limiting the present invention.

In another embodiment, a curve is determined according to a first user score and a fitting coefficient of default probability (or resource return probability, overdue probability) of the historical user, and the coefficient determination curve is used for determining a first calculation coefficient a of resource quota increasing amplitude (or decreasing amplitude) of the user.

Specifically, the coefficient determination curve is calculated using the following equation.

y＝Ax²+Bx+C (1)

Wherein y represents the utilization rate, overdue probability, default probability or resource return probability of the historical special resource voucher; x represents a first calculation coefficient a; a, B and C are curve parameters respectively.

It should be noted that, in the present invention, the statistical data (determined by the influence factors such as the usage rate of the specific resource voucher, the overdue probability, the maximum value, the minimum value, the average value of the default probability or the resource return probability, and the proportion of the default user or the overdue user to the user amount) of the different types of internet resource services in different historical time periods are set. Furthermore, the coefficient determination curve may also be a piecewise function of a plurality of line segments and/or curves over a particular time. The foregoing is illustrative only and is not to be construed as limiting the invention.

Therefore, a coefficient determination curve is fitted according to the first user score of the historical user, the utilization rate and default probability of the special resource voucher, the default probability or the resource return probability, and a more accurate coefficient determination curve can be obtained.

Specifically, according to the user feature data of the candidate users and the corresponding interconnection resource service types, the coefficient determining curve of each candidate user is selected, and the first calculating coefficient a of each candidate user is calculated by using the selected coefficient determining curve.

Specifically, a certain number of candidate users are screened out from the candidate user set (target user group) by multiplying the calculated first user score by the corresponding first calculation coefficient a to a final first user score and according to the final first user score.

In an embodiment, when the calculated first user score of the candidate user is greater than a set threshold, the candidate user is screened out, and the candidate user is, for example, a user whose internet resource quota is to be increased.

In the present embodiment, the set threshold is a value calculated from statistics of historical users over a certain period of time, but is not limited thereto, and may be a product of an average value of all statistical data and a corresponding first calculation coefficient. The foregoing is illustrative only and is not to be construed as limiting the invention.

Further, user feature data of the screened candidate users, such as user accounts, is obtained.

It should be noted that the foregoing is only an example for description, and the limitation of the present invention is not understood, and in other embodiments, the user feature data may also be a mobile phone number, a user equipment ID, or the like.

And then, inputting the user characteristic data of the candidate users into a trained resource quota determining model, calculating the user evaluation value of each candidate user, and determining a resource quota value or a resource quota range corresponding to the user evaluation value of the candidate user according to the calculated user evaluation value.

In yet another embodiment, the resource quota determination model directly outputs a resource quota value or a resource quota range for the candidate user.

Therefore, the user characteristic data of the candidate user is input into the trained resource quota determination model, and the user evaluation value is calculated, so that the resource quota value or the resource quota range needing to be adjusted can be accurately determined.

It should be noted that the foregoing is only described as an example, and the limitation of the present invention is not understood, and in other embodiments, the candidate user may also be a user whose internet resource quota is to be reduced.

In another example, as shown in fig. 3, a step S301 of presetting a plurality of resource quota intervals is further included.

In step S301, a plurality of resource quota intervals are set in advance for determining preset resource quota intervals corresponding to different user evaluation values.

In a specific embodiment, the resource quota interval is divided into a plurality of resource quota intervals by taking any value from 0.01 to 0.10 as an interval. For example, at 0.02 intervals, user ratings of 0.98-1.00 correspond to 8000-10000, 0.94-0.97 correspond to 6000-8000, and so on.

Specifically, according to the calculated user evaluation value, a corresponding preset resource quota interval is determined, and the resource quota of the candidate user is determined again.

For example, according to the calculated user evaluation value of 0.98, the corresponding preset resource quota is determined to be 8000-10000 yuan, and then according to the current resource quota of the candidate user, the resource quota which needs to be increased is calculated, so that the resource quota of the candidate user is determined again.

In the example shown in fig. 3, the method further includes step S301 of determining an adjustment coefficient of the resource quota to re-determine the resource quota of the candidate user.

In step S301, an adjustment coefficient of the resource quota is determined, so as to re-determine the resource quota of the candidate user according to the adjustment coefficient.

Specifically, the determining of the adjustment coefficient of the resource quota includes determining a second calculation coefficient, where the second calculation coefficient corresponding to the user evaluation value is determined according to the calculated user evaluation value.

Optionally, statistical calculation is performed on historical users in a specific time period, for example, a historical specific time period obtained by calculating 12 months ahead from a current time point, a preset resource quota interval with the largest number of users is selected, or a resource quota interval with the number of users greater than a set value or the top five of the ranking of the number of users (ranking from large to small), and the selected resource quota interval is a reference resource quota interval to be used for determining a corresponding second calculation coefficient.

Further, whether the user evaluation value of the candidate user is in a range corresponding to a reference resource quota interval is judged, and if the user evaluation value of the candidate user is in the range corresponding to the reference resource quota interval, a second calculation coefficient corresponding to the candidate user is determined; and if the resource quota is not in the range corresponding to the reference resource quota interval, determining a corresponding pre-resource quota interval by using the calculated user evaluation value so as to further determine the resource quota to be increased (or adjusted), thereby re-determining the resource quota of the candidate user.

For the adjustment coefficient, only the second calculation coefficient may be used as an adjustment coefficient of the resource quota, and a product of the first calculation coefficient a and the second calculation coefficient b may also be used as an adjustment coefficient for calculating the resource quota of the candidate user.

Further, the determined adjustment coefficient is multiplied by the determined resource quota interval (i.e., the resource quota interval corresponding to each candidate user), and the resource quota of the candidate user is re-determined, so that the resource quota of the candidate user can be determined more accurately.

Therefore, the resource quota of the candidate user can be redetermined according to the demand change of each user on the internet resource service along with the time change by calculating the adjustment coefficient of the determined resource quota and according to the adjustment coefficient and the determined resource quota interval, the resource quota can be more accurately determined, the user resource quota determining method can be further optimized, the distribution process of the internet resources can be further optimized, and the recyclability and the reusability of the internet resources can be effectively improved.

The procedures of the above-described method are merely for illustrating the present invention, and the order and number of the steps are not particularly limited. In addition, the steps in the method can be split into two or three steps, or some steps can be combined into one step, and the steps are adjusted according to practical examples.

Those skilled in the art will appreciate that all or part of the steps to implement the above-described embodiments are implemented as programs (computer programs) executed by a computer data processing apparatus. When the computer program is executed, the method provided by the invention can be realized. Furthermore, the computer program may be stored in a computer readable storage medium, which may be a readable storage medium such as a magnetic disk, an optical disk, a ROM, a RAM, or a storage array composed of a plurality of storage media, such as a magnetic disk or a magnetic tape storage array. The storage medium is not limited to centralized storage, but may be distributed storage, such as cloud storage based on cloud computing.

Embodiments of the apparatus of the present invention are described below, which may be used to perform method embodiments of the present invention. The details described in the device embodiments of the invention should be regarded as complementary to the above-described method embodiments; reference is made to the above-described method embodiments for details not disclosed in the apparatus embodiments of the invention.

Example 2

Referring to fig. 4, 5 and 6, the present invention further provides a resource quota re-determining apparatus 400 based on a user tag, where the resource quota re-determining apparatus 400 includes: the screening module 401 is configured to determine a reference user and user characteristics thereof, and perform clustering calculation on users in the candidate user set and the reference user to screen out a target user group; an establishing module 402, which regenerates a user label based on the user characteristics of the screened target user group, and establishes a training data set for training a resource quota determining model according to the user label; and a calculation determining module 403, configured to input the user characteristics of the candidate user into the trained resource quota determining model, and output the resource quota corresponding to the candidate user.

As shown in fig. 5, the apparatus 400 for re-determining resource quota further includes a determining module 501, that is, the filtering module 401 is divided into the determining module 501 and the filtering module 401. The determining module 501 is configured to determine a reference user and user characteristics thereof, and specifically, is configured to determine a reference user having a resource quota value greater than that of a candidate user in the candidate user set. And the system is also used for determining the user basic information characteristics, the professional information characteristics and the resource return performance characteristics of the reference user.

Further, the users in the candidate user set and the reference user are subjected to clustering calculation so as to screen out a target user group.

Optionally, a K-means algorithm is used to perform clustering calculation or multiple rounds of clustering calculation on the users in the candidate user set and the reference user. Generating a plurality of user groups after the clustering calculation, and screening a target user group from the plurality of user groups according to a screening rule; wherein the screening rule includes that the total number of users in the user group is greater than a certain number, the ratio of the number of the reference users to the number of the candidate users in the user group is within a predetermined proportion range, and the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is compared.

In one embodiment, when the total number of users in the user group is greater than a certain number and the ratio of the number of the reference users to the number of the candidate users in the user group is 1:1, the user group is determined to be a target user group.

As shown in fig. 6, the resource quota re-determining apparatus 400 further includes a generating module 601, where the generating module 601 generates a user tag based on the user characteristics of the screened target user group.

Specifically, the generating module 601 regenerates the user tag according to the user characteristics of the screened target user group, where the screened user is a user who needs to adjust (increase or decrease) the resource quota.

In another embodiment, the generating module 601 redefines the user label that is newly changed to 1 for the user whose similarity to the candidate user in the target user group is greater than a specified value (e.g., the specified value is 70%, 80%, or 90%) and redefines the positive sample and the negative sample for the user whose similarity to the candidate user in the target user group is less than or equal to the specified value (e.g., the specified value is 70%, 80%, or 90%) to be changed to 0, so as to establish the training data set.

Optionally, the resource quota determination model is constructed using a logistic regression model, an Xgboost model, and/or a deep neural network.

Specifically, the resource quota determining model is trained by using the established training data set, the trained resource quota determining model is used for calculating a user evaluation value, a resource quota value or a resource quota range corresponding to the user evaluation value of the candidate user is output, and the user evaluation value is a numerical value between 0 and 1.

And further, inputting the user characteristics of the candidate users into the trained resource quota determining model, and outputting the resource quotas corresponding to the candidate users.

Specifically, user characteristics of a candidate user are obtained, and the user characteristics of the candidate user are input into a trained resource quota determination model to calculate a user evaluation value of the candidate user.

Further, the method further comprises determining an adjustment coefficient of the resource quota, and re-determining the resource quota of the candidate user according to the adjustment coefficient.

In embodiment 2, the same portions as those in embodiment 1 are not described.

Those skilled in the art will appreciate that the modules in the above-described embodiments of the apparatus may be distributed as described in the apparatus, and may be correspondingly modified and distributed in one or more apparatuses other than the above-described embodiments. The modules of the above embodiments may be combined into one module, or further split into multiple sub-modules.

Example 3

In the following, embodiments of the computer apparatus of the present invention are described, which may be seen as specific physical embodiments for the above-described embodiments of the method and apparatus of the present invention. The details described in the computer device embodiment of the invention should be considered as additions to the method or apparatus embodiment described above; for details which are not disclosed in the embodiments of the computer device of the invention, reference may be made to the above-described embodiments of the method or apparatus.

Fig. 7 is a block diagram of an exemplary embodiment of a computer device according to the present invention. A computer apparatus 200 according to this embodiment of the present invention is described below with reference to fig. 7. The computer device 200 shown in fig. 7 is only an example and should not bring any limitation to the function and the scope of use of the embodiments of the present invention.

As shown in FIG. 7, computer device 200 is in the form of a general purpose computing device. The components of computer device 200 may include, but are not limited to: at least one processing unit 210, at least one storage unit 220, a bus 230 connecting different device components (including the storage unit 220 and the processing unit 210), a display unit 240, and the like.

Wherein the storage unit stores program code executable by the processing unit 210 to cause the processing unit 210 to perform steps according to various exemplary embodiments of the present invention described in the processing method section of the above-mentioned computer apparatus of the present specification. For example, the processing unit 210 may perform the steps as shown in fig. 1.

The memory unit 220 may include readable media in the form of volatile memory units, such as a random access memory unit (RAM)2201 and/or a cache memory unit 2202, and may further include a read only memory unit (ROM) 2203.

The storage unit 220 may also include a program/utility 2204 having a set (at least one) of program modules 2205, such program modules 2205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.

Bus 230 may be one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

The computer device 200 may also communicate with one or more external devices 300 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the computer device 200, and/or with any devices (e.g., router, modem, etc.) that enable the computer device 200 to communicate with one or more other computing devices. Such communication may occur via an input/output (I/O) interface 250. Also, computer device 200 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the Internet) through network adapter 260. Network adapter 260 may communicate with other modules of computer device 200 via bus 230. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the computer device 200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments of the present invention described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiment of the present invention can be embodied in the form of a software product, which can be stored in a computer-readable storage medium (which can be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to make a computing device (which can be a personal computer, a server, or a network device, etc.) execute the above-mentioned method according to the present invention. Which when executed by a data processing device, enables the computer program product to carry out the above-mentioned method of the invention.

As shown in fig. 8, the computer program may be stored on one or more computer program products. The computer program product may be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer program product include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The computer readable storage medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer program product may be transmitted, propagated, or transported for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).

In summary, the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that some or all of the functionality of some or all of the components in embodiments in accordance with the invention may be implemented in practice using a general purpose data processing device such as a microprocessor or a Digital Signal Processor (DSP). The present invention may also be embodied as apparatus or device programs (e.g., computer programs and computer program products) for performing a portion or all of the methods described herein. Such a program implementing the invention may be stored on a computer program product or may be in the form of one or more signals. Such a signal may be downloaded from an internet website or provided on a carrier signal or in any other form.

While the foregoing embodiments have described the objects, aspects and advantages of the present invention in further detail, it should be understood that the present invention is not inherently related to any particular computer, virtual machine or electronic device, and various general-purpose machines may be used to implement the present invention. The invention is not to be considered as limited to the specific embodiments thereof, but is to be understood as being modified in all respects, all changes and equivalents that come within the spirit and scope of the invention.

Claims

1. A resource quota re-determining method based on a user tag is characterized by comprising the following steps:

determining a reference user and user characteristics thereof, and performing cluster calculation on the users in the candidate user set and the reference user to screen out a target user group;

regenerating a user label based on the user characteristics of the screened target user group, and establishing a training data set for training a resource quota determination model according to the user label;

inputting the user characteristics of the candidate users into the trained resource quota determining model, and outputting the resource quota corresponding to the candidate users.

2. The method according to claim 1, wherein the determining the reference user and the user characteristics thereof comprises:

determining a reference user with a resource quota value larger than that of the candidate user in the candidate user set; and the number of the first and second electrodes,

and determining the user basic information characteristics, the professional information characteristics and the resource return performance characteristics of the reference user.

3. The method according to claim 1 or 2, wherein the clustering the users in the candidate user set with the reference user to filter out a target user group comprises:

generating a plurality of user groups after the clustering calculation, and screening a target user group from the plurality of user groups according to a screening rule; wherein the content of the first and second substances,

the screening rule includes that the total number of users in the user group is greater than a certain number, the ratio of the number of the reference users to the number of the candidate users in the user group is within a predetermined proportion range, and the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is compared.

4. The resource quota re-determining method of claim 3, wherein,

and when the total number of the users of the user group is greater than a specific number and the ratio of the number of the reference users to the number of the candidate users in the user group is 1:1, determining that the user group is a target user group.

5. The resource quota re-determining method of claim 3, wherein,

and when the total number of the users in the user group is greater than a specific number, the ratio of the number of the reference users to the number of the candidate users in the user group is 1:1, and/or the similarity between the user characteristics of the reference users and the user characteristics of the candidate users in the user group is greater than a set value, determining that the user group is a target user group.

6. The method of claim 1, wherein the generating the user tag based on the user characteristics of the screened target user group comprises:

redefining user labels which are similar to the candidate users in the target user group and become 1, and redefining user labels which are dissimilar to the candidate users in the target user group and become 0 to establish a training data set; and/or

And redefining the positive sample and the negative sample to the user label of which the similarity with the candidate user in the target user group is more than the specified value and the user label of which the similarity with the candidate user in the target user group is less than or equal to the specified value and the user label of which the similarity with the candidate user in the target user group is 0 so as to establish a training data set.

7. The resource quota re-determining method of claim 1,

and performing clustering calculation or multi-round clustering calculation on the users in the candidate user set and the reference user by using a K-means algorithm.

8. The method of claim 1, wherein constructing the resource quota determination model comprises:

establishing a resource quota determining model by using a logistic regression model, an Xgboost model and/or a deep neural network, wherein the resource quota determining model calculates a user evaluation value, and outputs a resource quota value or a resource quota range corresponding to the user evaluation value of the candidate user, and the user evaluation value is a numerical value between 0 and 1; or

And the resource quota determining model directly outputs the resource quota value or the resource quota range of the candidate user.

9. A resource quota re-determination apparatus based on a user tag, comprising:

the screening module is used for determining reference users and user characteristics thereof, and performing clustering calculation on the users in the candidate user set and the reference users so as to screen out a target user group;

the establishing module is used for regenerating a user label based on the user characteristics of the screened target user group and establishing a training data set for training a resource quota determining model according to the user label;

and the calculation determining module is used for inputting the user characteristics of the candidate users into the trained resource quota determining model and outputting the resource quotas corresponding to the candidate users.

10. A computer device comprising a processor and a memory for storing a computer executable program, the processor performing the user tag-based resource quota re-determining method of claim 1 when the computer program is executed by the processor.