CN116049560A

CN116049560A - Information processing method and device

Info

Publication number: CN116049560A
Application number: CN202310122545.9A
Authority: CN
Inventors: 孙冉
Original assignee: Alibaba China Co Ltd
Current assignee: Alibaba China Co Ltd
Priority date: 2023-01-19
Filing date: 2023-01-19
Publication date: 2023-05-02

Abstract

The embodiment of the specification provides an information processing method and device, wherein the information processing method comprises the following steps: selecting users with unchanged user grades from the user set to construct a non-transition user group; according to the service participation information of the non-transition users in the target service in the non-transition user group, selecting users to be transitioned with user grades to be changed from the non-transition user group; inputting the interaction information of the user to be transitioned in the preset interaction dimension of the target service to a target attribution model for processing to obtain target interaction information of the user to be transitioned; and creating a grade change strategy associated with the user to be transitioned based on the target interaction information.

Description

Information processing method and device

Technical Field

The embodiment of the specification relates to the technical field of computers, in particular to an information processing method.

Background

With the development of internet technology and the gradual maturity of an e-commerce platform, the consumer shopping on the e-commerce platform becomes a part of daily life. However, as the internet traffic becomes saturated, the value of the merchant cannot be effectively improved by the merchant in the current e-commerce field through a new traffic increasing mode. And thus is particularly important for the fine operation of existing subscribers. In the prior art, the merchant value is generally improved by adopting modes of developing new users, pushing activities for the new and old users and the like. However, the manner of unified pushing activity for all users cannot meet the personalized requirements of the users, the influence on the grade change of the users is small, and the accurate access of the users cannot be realized. Therefore, an information processing method is demanded to solve the above-described problems.

Disclosure of Invention

In view of this, the present embodiment provides an information processing method. One or more embodiments of the present specification also relate to an information processing apparatus, a computing device, a computer-readable storage medium, and a computer program that solve the technical drawbacks existing in the prior art.

According to a first aspect of embodiments of the present specification, there is provided an information processing method including:

selecting users with unchanged user grades from the user set to construct a non-transition user group;

according to the service participation information of the non-transition users in the target service in the non-transition user group, selecting users to be transitioned with user grades to be changed from the non-transition user group;

inputting the interaction information of the user to be transitioned in the preset interaction dimension of the target service to a target attribution model for processing to obtain target interaction information of the user to be transitioned;

and creating a grade change strategy associated with the user to be transitioned based on the target interaction information.

According to a second aspect of the embodiments of the present specification, there is provided an information processing apparatus including:

a construction module configured to select a user whose user level is unchanged from the user set to construct a non-transition user group;

The screening module is configured to screen users to be transited, the user grade of which is to be changed, from the non-transition user group according to the service participation information of the non-transition users in the target service;

the input module is configured to input the interaction information of the user to be transitioned in the preset interaction dimension of the target service to a target attribution model for processing to obtain target interaction information of the user to be transitioned;

and the creation module is configured to create a grade change strategy associated with the user to be transitioned based on the target interaction information.

According to a third aspect of embodiments of the present specification, there is provided a computing device comprising:

a memory and a processor;

the memory is configured to store computer-executable instructions, and the processor is configured to execute the computer-executable instructions, which when executed by the processor, implement the steps of the information processing method described above.

According to a fourth aspect of the embodiments of the present specification, there is provided a computer-readable storage medium storing computer-executable instructions which, when executed by a processor, implement the steps of the above-described information processing method.

According to a fifth aspect of the embodiments of the present specification, there is provided a computer program, wherein the computer program, when executed in a computer, causes the computer to perform the steps of the above-described information processing method.

An information processing method provided in one embodiment of the present disclosure constructs a non-transition user group by selecting a user whose user class is unchanged from a user set; according to the service participation information of the non-transition users in the target service in the non-transition user group, selecting users to be transitioned with user grades to be changed from the non-transition user group; inputting the interaction information of the user to be transitioned in the preset interaction dimension of the target service to a target attribution model for processing to obtain target interaction information of the user to be transitioned; and creating a grade change strategy associated with the user to be transitioned based on the target interaction information. According to the business participation information of the non-transition users, the users to be transited with the user grade to be changed are screened from the non-transition user group, the accuracy of user screening can be improved, the interaction information is processed through the attribution model, the target interaction information is determined in the interaction information, the accuracy of target interaction information determination can be improved, and then a grade change strategy corresponding to the target business information is created, so that the grade change of the users to be transited is promoted, and the accurate touch of the users is realized.

Drawings

Fig. 1 is a schematic structural diagram of an information processing method according to an embodiment of the present application;

FIG. 2 is a flow chart of a method of information processing provided in one embodiment of the present disclosure;

FIG. 3 is a process flow diagram of a method of information processing provided in one embodiment of the present disclosure;

FIG. 4 is a flowchart of a processing procedure of an information processing method according to an embodiment of the present disclosure;

fig. 5 is a schematic structural view of an information processing apparatus provided in one embodiment of the present specification;

FIG. 6 is a block diagram of a computing device provided in one embodiment of the present description.

Detailed Description

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present description. This description may be embodied in many other forms than described herein and similarly generalized by those skilled in the art to whom this disclosure pertains without departing from the spirit of the disclosure and, therefore, this disclosure is not limited by the specific implementations disclosed below.

The terminology used in the one or more embodiments of the specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of the specification. As used in this specification, one or more embodiments and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used in one or more embodiments of the present specification refers to and encompasses any or all possible combinations of one or more of the associated listed items.

It should be understood that, although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, a first may also be referred to as a second, and similarly, a second may also be referred to as a first, without departing from the scope of one or more embodiments of the present description. The word "if" as used herein may be interpreted as "at … …" or "at … …" or "responsive to a determination", depending on the context.

First, terms related to one or more embodiments of the present specification will be explained.

MILs: multiple instance learning (multiple-instance learning) is a learning problem with multiple instance packages (bag) as training units. In multi-instance learning, the training set consists of a set of multi-instance packages (bag) with classification labels, each multi-package (bag) containing several instances without classification labels. If the multiple instance package (bag) contains at least one positive instance (instance), then the package is marked as a positive-class multiple instance package (positive package). If all examples of the multi-example package are negative examples, the package is marked as a negative-class multi-example package (negative package). The purpose of multi-instance learning is to build a multi-instance classifier by learning on multi-instance packets with classification labels and apply the classifier to the predictions of unknown multi-instance packets.

With the saturation of internet traffic, resources obtained by merchants in the current e-commerce field in a newly-added traffic mode gradually reach the bottleneck, so that the refined operation of the existing users becomes particularly important. The existing user group can be divided into a high-value layer and a low-value layer according to the gmv (commodity transaction total) contributed by the user group, and if the conversion time between the high-value layer and the low-value layer of a certain user can be predicted and the conversion reason can be predicted, the loss of the user from the high-value layer to the low-value layer can be effectively prevented, or the conversion from the low-value layer to the high-value layer can be promoted, so that the high-value user quantity of a merchant can be promoted. Based on the scheme, the scheme for improving the high-value user quantity and further improving the value of the merchant is provided. Users who are likely to undergo state transition between the high value and the low value are predicted through the algorithm model, and conversion reasons of the users are predicted in a personalized mode, so that the value of merchants is improved. The transition crowd between the potential high-value and low-value levels is predicted by a transition prediction model, key reasons influencing the transition are mined by a multi-example learning algorithm, and personalized intervention strategies are adopted for the potential transition crowd based on the key reasons, so that the high-value user quantity of the merchant is stabilized at a higher level, and the value of the merchant also tends to be stabilized.

Fig. 1 is a schematic structural diagram of an information processing method according to an embodiment of the present application, as shown in fig. 1, based on a user group formed by users browsing a store in a period of time, a non-transition user group is formed by selecting users whose user level is unchanged in the period of time from the user group. And screening the users to be transited, the user grade of which is to be changed, from the non-transition user group according to the service participation information of the non-transition users in the target service. And inputting the interaction information of the user to be transitioned in the preset interaction dimension of the target service to the target attribution model for processing, outputting the target interaction information of the user to be transitioned, and further creating a grade change strategy associated with the user to be transitioned based on the target interaction information. The method is convenient for determining corresponding pushing tasks or activities based on the level change strategy of each user to be transitioned, pushing the tasks or activities to the user to be transitioned, and realizing personalized task recommendation, so that the high-value user quantity of the merchant is stabilized at a higher level, and the value of the merchant also tends to be stabilized at a level.

In the present specification, an information processing method is provided, and the present specification relates to an information processing apparatus, a computing device, a computer-readable storage medium, and a computer program, one by one, in the following embodiments.

Referring to fig. 2, fig. 2 shows a flowchart of an information processing method according to an embodiment of the present disclosure, which specifically includes the following steps.

Step S202: and selecting users with unchanged user grades from the user set to construct a non-transition user group.

Specifically, the user set refers to a whole formed by users, or a certain space for storing the users, the users correspond to different types of people under different scenes, in the field of e-commerce, the consumers are users, and the users can realize operations such as browsing, purchasing and the like on an e-commerce platform; a user may also be referred to as a user, such as use of an application, use of a network, use of physical items, etc.; the user level refers to a level of user classification according to a certain rule, for example, classification according to the use time of the user, classification according to the resource occupation amount of the user, and the like; the transition means a change of the level, the upward transition means an increase of the level, the downward transition means a decrease of the level, and the non-transition user group means a whole composed of non-transition users, and the non-transition users means users who have not undergone the level transition within a fixed time range, for example, users who have not undergone the level change within seven days in the user group.

Based on the method, a user set composed of at least one user is obtained, users with unchanged user grades in a preset time range are selected from the user set to serve as non-transition users, and a non-transition user group is composed based on the selected non-transition users.

In practical application, under the electronic market scene, the users in the user set can be all users who browse a certain store, or all users who exchange resources for a certain commodity in the store; in the case where the user uses the application program with the service function, the user in the user set may be the user who completes registration in the application program, or may be the user of the application program, which is not limited in this embodiment.

Further, considering that the behavior habits of users are different and the demands are different, when selecting users with unchanged user grades, transition information of the users within a period of time is also required to be considered, and the method is specifically implemented as follows:

reading historical transition information of users in the user set in a preset time interval; selecting non-transition users with unchanged user grades in the user set based on the historical transition information; and constructing a non-transition user group based on the non-transition user.

Specifically, the preset time interval refers to a preset time range, and the units of the time interval can be hours, days, weeks, months and the like; the historical transition information refers to grade change information of a user, and the historical transition information comprises, but is not limited to, transition time information, transition grade information, behavior information affecting transition and the like, wherein the transition grade information comprises grade information before transition and grade information after transition, for example, the user transitions from grade 1 to grade 2; the transition behavior information refers to the behavior of the user with respect to the store or the commodity before the user makes the level change, the behavior with respect to the store includes but is not limited to browsing, collecting, sharing, etc., and the behavior with respect to the commodity includes but is not limited to browsing, collecting, purchasing, sharing, removing, etc.

Based on the above, after the user set is obtained, the historical transition information of each user in the user set in a preset time interval is read, whether the user has level change in the preset time interval is determined according to the historical transition information of the user, and the user which has no level change in the preset time interval in the user set is taken as a non-transition user. And forming a non-transition user group by the screened non-transition users, wherein the grade change comprises grade improvement and grade reduction.

For example, under the service application program and the electronic market, the user can browse shops and purchase goods in an online manner, the merchant can customize a user grade system, grade the user based on the grade system, and the actions of browsing, purchasing, sharing, collecting and the like of the goods of the user can influence the grade division of the user. The user's level may be classified according to the total Life Time Value (LTV) of the user. And acquiring a user set, wherein users which do not generate level change within three days are selected as non-transition users according to the transition information of each user in the user set, and the non-transition users form a non-transition user group. To facilitate subsequent selection of the user to be transitioned among the group of non-transition users.

In summary, by reading the historical transition information of the users in the user set, and further screening the non-transition users with unchanged user grades in the user set, the screening accuracy of the non-transition users can be improved.

Step S204: and screening the users to be transited, the user grade of which is to be changed, from the non-transition user group according to the service participation information of the non-transition users in the target service.

Specifically, after the non-transition user group is constructed by selecting the user with unchanged user level in the user set, the user to be transitioned with the user level to be changed can be screened in the non-transition user group according to the service participation information of the non-transition user in the target service, wherein the service participation information refers to the behavior information generated by the user participating in the target service, and the behavior information under the electronic market scene comprises but is not limited to browsing information, collection information, sharing information and activity participation information of shops, and browsing information, collection information, sharing information, purchase information and the like of commodities; the user to be transitioned refers to a user which predicts according to the business participation information and the predicted level may change, namely, the user to be transitioned refers to a user with a level change trend.

Based on the information, after the user with unchanged user grade is selected from the user set to construct a non-transition user group, the business participation information of each non-transition user in the non-transition user group in the target business is read, and the user with unchanged user grade is selected from the non-transition user group as the user to be transitioned according to the business participation information of each non-transition user. The determining of the business participation information comprises the following steps: reading resource participation information, operation information and/or browsing information of a non-transition user in the target service in the non-transition user group; and taking the resource participation information, the operation information and/or the browsing information as service participation information. The resource engagement information includes, but is not limited to, purchase information for the merchandise; the operation information includes, but is not limited to, browsing information, collection information, sharing information, purchase information, etc.; business participation information includes, but is not limited to, activity participation information, commodity purchase information, and the like.

In practical application, because the target services participated by the non-transition users are different, the generated service participation information is also different, and when the target service is shopping, the service participation information comprises but is not limited to information such as purchasing, collecting, sharing and browsing of commodities, and information such as collecting, sharing and browsing of shops; when the target service is used by the application program, the service participation information includes, but is not limited to, information such as application program login information, online time length and the like.

Further, in order to improve screening accuracy and screening efficiency of the user to be transitioned, screening may be performed by inputting the screening result into a neural network model, which is specifically implemented as follows:

reading service participation information of non-transition users in the non-transition user group in the target service, and inputting the service participation information into a target prediction model for prediction; and screening the users to be changed in the user level in the non-transition user group according to the prediction result.

Specifically, the target prediction model is a neural network model for user screening based on service participation information of a non-transition user in target service, training of the prediction model can be completed in a model training mode, and an initial prediction model is trained into a target prediction model; the prediction result is a result obtained by predicting each non-transition user, and the prediction result comprises transition and non-transition, namely, a label is allocated to each non-transition user, a 'transition' label is allocated to a user with higher possibility of transition, the user is taken as a user to be transitioned, and correspondingly, a 'non-transition' label is allocated to a user with lower possibility of transition.

In practical application, when the business participation information of each non-transition user is input into the target prediction model for prediction, the prediction result of the model can also be a prediction score obtained by scoring each non-transition user, and the users to be transitioned are screened from the non-transition users according to the prediction score.

Based on the information, the service participation information of the non-transition users in the target service in the non-transition user group is read, the service participation information of each non-transition user is input into a target prediction model to be predicted, a prediction result of each non-transition user is obtained, and users to be changed in user grade are screened from the non-transition user group according to the prediction result.

Along the above example, the user to be transitioned is selected from the group of non-transition users consisting of 100 users. Inputting business participation information such as commodity purchase, sharing and collection corresponding to each user in 100 users, store browsing and activity participation into a trained target prediction model for prediction to obtain a prediction result of each user, wherein the target prediction model can predict the users as users to be transited and common users, and the users to be transited are users with high transition possibility.

In summary, the service participation information of the non-transition users in the non-transition user group in the target service is input into the target detection model for prediction, so that the users to be transitioned are screened in the non-transition user group, and the prediction accuracy of the users to be transitioned can be improved.

Further, in order to ensure the prediction accuracy of the target prediction model, the initial prediction model may be trained based on the training sample, and the target prediction model meeting the training stop condition is known to be obtained, which is specifically implemented as follows:

selecting an upward transition user from a training user set, and taking transition participation information of the upward transition user as a first positive sample; selecting a downward transition user from the training user set, taking transition participation information of the downward transition user as a second positive sub-sample, and forming a positive sample by the first positive sub-sample and the second positive sub-sample; performing negative sampling on the training user set to obtain a negative sample; and training the initial prediction model based on the positive sample and the negative sample until a target prediction model meeting the training stop condition is obtained.

Specifically, the training user set refers to a user set composed of users for model training; the upward transition user is a user with the level increased based on the original level of the user; correspondingly, the downward transition user refers to a user with the grade reduced based on the original grade of the user; the transition participation information refers to level change information, i.e., level up information or level down information; the level increasing information is a first positive sub-sample, the level decreasing information is a second positive sub-sample, the positive sample is formed by the first positive sub-sample and the second positive sub-sample, and the positive sample is the information of the level transition; the negative sampling means that a negative sample is selected from the training user set, random sampling is carried out on training users which do not have level change in the training user set, and the obtained transition participation information corresponding to the training users is the negative sample.

Based on the information, when training a prediction model, selecting an upward transition user from a training user set, and taking transition participation information of the upward transition user as a first positive sub-sample; selecting a downward transition user from the training user set, taking transition participation information of the downward transition user as a second positive sample, wherein the first positive sample and the second positive sample form a positive sample for model training; and carrying out negative sampling on training users which do not have grade change in the training user set, determining transition participation information of the selected training users according to sampling results, and taking the transition participation information as a negative sample for model training. And training the initial prediction model based on the positive sample and the negative sample until a target prediction model meeting the training stop condition is obtained.

Along the above example, a user associated with a store within one month is selected as a training user, and a training user set is composed of training users, wherein the users associated with the store include, but are not limited to, users who browse, purchase, collect goods in the store, and users who browse through the store. Reading the grade change information of the users in the training user set within one month, and taking the grade change information corresponding to the users with the improved grade as a first positive sub-sample; taking the grade change information corresponding to the user with the reduced grade as a second positive sub-sample; and sampling among users with unchanged grades in the training user set, and taking grade change information corresponding to the sampled users as a negative sample. And training the initial prediction model according to the positive and negative samples until a target prediction model meeting training stop conditions is obtained, wherein the training stop conditions include, but are not limited to, the number of training times for completing a set number, the prediction accuracy of the model reaching a set value, and the like.

In summary, the positive sample and the negative sample are obtained by sampling in the training user set, and then model training is performed based on the positive sample and the negative sample, so that the training effect of the model is improved.

Further, considering that the user level change includes two kinds of upward change and downward change, that is, there may be an increase in user level and two kinds of decrease in user level, since it is difficult to complete prediction of two kinds of different user level changes by using one prediction model, two initial prediction models may be constructed to predict two kinds of level change conditions respectively, which is specifically implemented as follows:

constructing a user map based on training user information of training users in the training user set; constructing an initial vector of a user node in the user map; a first initial prediction model and a second initial prediction model are generated based on the initial vector of each training user, and the first initial prediction model and the second initial prediction model are used as initial prediction models.

Specifically, the training user information refers to user information and object information associated with the training user, wherein the user information includes, but is not limited to, user labels, user ids, commodity information associated with the user, merchant information and the like; the object information includes, but is not limited to, information such as commodity id, merchant id, commodity attribute, etc.; the user profile refers to a profile constructed based on training user information, wherein the determining of the user profile comprises: reading training user information of training users in the training user set; constructing a user map by taking the user information and the service information in the training user information as nodes and taking the relation between the user information and the service information as sides; the user nodes comprise all training users, and initial vectors corresponding to the user nodes in the user map are constructed, and the initial vectors of each training user in the user nodes are actually constructed; the first initial predictive model and the second initial predictive model are the same as the models used to make the user level predictions, except that the first initial predictive model may be trained to predict users with increased levels; accordingly, the second initial predictive model may be trained to predict users with decreasing levels.

Based on the training user information of the training users in the training user set, a user map is constructed, wherein the training user information of the training users in the training user set can be read when the user map is constructed, the user information and the service information in the training user information are taken as nodes, and the relationship between the user information and the service information is taken as a side to construct the user map. Training users contained in user nodes of the user graph are determined, and an initial vector for each training user is constructed. A first initial prediction model and a second initial prediction model are generated based on the initial vector of each training user, and the first initial prediction model and the second initial prediction model are used as initial prediction models.

Along the above example, a training user set composed of 2000 training users is obtained, and information such as commodity purchase, browsing, sharing and the like, information such as store browsing, collection and the like, and activity participation information corresponding to each training user are determined. Constructing a user map shown in fig. 3 (a) by taking training user set information and commodity information as nodes. The user map is constructed by taking user id, user information, commodities, commodity information, merchant information and the like as nodes and taking the relation between the nodes as sides. And constructing an initial vector of each training user in the user nodes of the user map by adopting a graph learning algorithm, so as to realize the pre-training of the user nodes. And constructing a first initial prediction model and a second initial prediction model according to the initial vector of each training user, and taking the first initial prediction model and the second initial prediction model as initial prediction models.

In summary, the user map is constructed based on the training user information of the training users in the training user set, so as to generate a first initial prediction model and a second initial prediction model, so that the first initial prediction model is trained into a prediction model with the capability of the user with the increased recognition level, and the second initial prediction model is trained into a prediction model with the capability of the user with the reduced recognition level.

Further, considering that the change of the user level includes upward transition and downward transition, in order to more accurately identify the user with upward transition and the user with downward transition, the method can be implemented by training two prediction models with different prediction capabilities, and specifically is implemented as follows:

training the first initial prediction model based on the first positive sub-sample and the negative sub-sample until a first target prediction model meeting a training stop condition is obtained; training the second initial prediction model based on the second positive sub-sample and the negative sub-sample until a second target prediction model meeting a training stop condition is obtained; and taking the first target prediction model and the second target prediction model as target prediction models.

Based on the first target prediction model, taking the first positive sub-sample and the negative sub-sample as training samples, and training the first initial prediction model until a first target prediction model meeting training stop conditions is obtained; training a second initial prediction model by taking the second positive sub-sample and the negative sample as training samples until a second target prediction model meeting training stop conditions is obtained; and taking the first target prediction model and the second target prediction model as target prediction models.

Along the use example, reading the grade change information of the user in the training user set within one month, taking the grade change information corresponding to the user with the improved grade as a first positive sub-sample, and training a first initial prediction model by combining the negative sample, so that the trained first target prediction model has the capability of identifying the user with the improved grade; and taking the grade change information corresponding to the user with the reduced grade as a second positive sub-sample, and training a second initial prediction model by combining the negative samples, so that the second target prediction model has the capability of identifying the user with the reduced grade.

In summary, the first target prediction model meeting the training stop condition is obtained by training the first initial prediction model based on the first positive sub-sample and the negative sample; the second initial prediction model is trained based on the second positive sub-sample and the negative sub-sample until a second target prediction model meeting the training stop condition is obtained, so that the first target prediction model has the capability of identifying users with improved grades, and the second prediction model has the capability of identifying users with reduced grades, and the accuracy of user grade change prediction is improved.

Further, considering that the non-transition user group includes users that may perform upward transition and users that may perform downward transition, in order to improve the determination efficiency of the users to be transitioned, the non-transition users may be simultaneously input into the first target prediction model and the second target prediction model to perform prediction, so as to obtain the users to be transitioned, which is specifically implemented as follows:

inputting service participation information of non-transition users in the non-transition user group in the target service into a first target prediction model and a second target prediction model for prediction; screening a first user to be transited, the user grade of which is to be changed, from the non-transition user group according to a first prediction result corresponding to the first target prediction model, and screening a second user to be transited, the user grade of which is to be changed, from the non-transition user group according to a second prediction result corresponding to the second target prediction model; and forming a user to be transitioned by the first user to be transitioned and the second user to be transitioned.

Specifically, the first prediction result corresponding to the first target prediction model refers to the prediction result of the first prediction model on each piece of service participation information, and correspondingly, the second prediction result corresponding to the second target prediction model refers to the prediction result of the second prediction model on each piece of service participation information, where the first prediction result may be whether a non-transition user is a user that may transition upwards, and correspondingly, the second prediction result may be whether the non-transition user is a user that may transition downwards, and the user that may transition upwards is the first transition user, and the user that may transition downwards is the second transition user.

Based on the information, the service participation information of the non-transition users in the non-transition user group in the target service is input into a first target prediction model for prediction, and a first user to be transitioned, the user grade of which is to be changed, is screened from the non-transition user group according to a first prediction result corresponding to the first target prediction model; and inputting service participation information of the non-transition users in the target service in the non-transition user group into a second target prediction model for prediction, and screening a second user to be transitioned, the user grade of which is to be changed, from the non-transition user group according to a second prediction result corresponding to the second target prediction model. And the first user to be transitioned and the second user to be transitioned form the user to be transitioned.

According to the method, the non-transition user group comprises 100 non-transition users, service participation information corresponding to the 100 non-transition users can be respectively input into a first target prediction model and a second target prediction model for prediction, 20 users which can be in upward transition are screened out through the first target prediction model, 25 users which can be in downward transition are screened out through the second target prediction model, and then the users to be in upward transition are composed of 20 users which can be in upward transition and 25 users which can be in downward transition.

In addition, the service participation information of the non-transition users in the target service in the non-transition user group can be input into a first target prediction model for prediction; screening a first user to be transitioned, the user grade of which is to be changed, from the non-transition user group according to a first prediction result; taking the users except the first user to be transited in the non-transition users as users to be predicted, and inputting the business participation information of the users to be predicted in the target business into a second target prediction model for prediction; and screening a second user to be transited, the user grade of which is to be changed, from the users to be predicted according to a second prediction result.

For example, the non-transition user group includes 100 non-transition users, the service participation information corresponding to the 100 non-transition users may be input to the first target prediction model to predict, 20 users possibly transitioning upwards are screened out, 80 users except for the 20 users determined to transition upwards in the 100 non-transition users are input to the second target prediction model to predict, 25 users possibly transitioning downwards are screened out again, and then the 20 users possibly transitioning upwards and the 25 users possibly transitioning downwards form the user to be transitioned.

In summary, the service participation information of the non-transition user in the target service is input to the first target prediction model and the second target prediction model to predict, so as to obtain the user to be transitioned in the non-transition user, thereby improving the determination efficiency and accuracy of the user to be transitioned.

Step S206: and inputting the interaction information of the user to be transitioned in the preset interaction dimension of the target service to a target attribution model for processing to obtain the target interaction information of the user to be transitioned.

Specifically, after the to-be-transitioned user with the user grade to be changed is screened in the non-transition user group according to the service participation information of the non-transition user in the target service, the interaction information of the to-be-transitioned user in the target service preset interaction dimension is input into the target attribution model for processing to obtain the target interaction information of the to-be-transitioned user, wherein the preset interaction dimension refers to the user behavior dimension associated with the target service, and the preset interaction dimension can be the activity dimension of the user participation or the commodity purchasing dimension of the user; correspondingly, the interaction information refers to information of the user to be transited in corresponding dimensions, including but not limited to the number of times of participating in the activity, the name or number of the participating activity, purchased goods and the like; the attribution model is a neural network model for processing the interaction information, and is used for scoring each interaction information, further determining target interaction information in the interaction information according to the score, and correspondingly, the target interaction information is at least one interaction information with a higher score screened in the interaction information.

Based on the information, after the user to be changed in the user grade is screened in the non-transition user group according to the service participation information of the non-transition user in the target service, the interaction information of the user to be changed in the target service preset interaction dimension is read, the interaction information of each user to be transitioned in the preset interaction dimension is input into the target attribution model for processing, the target interaction information of the user to be transitioned is obtained, and the target interaction information is selected in the interaction information of the target user. The determination of the interaction information comprises the following steps: reading resource interaction information, resource score information and/or task participation information of the user to be transitioned in the preset interaction dimension of the target service; and taking the resource interaction information, the resource score information and/or the task participation information as interaction information.

In practical application, the determination of the target interaction information can be performed in a multi-instance learning mode, the user interaction behavior corresponding to each interaction information can be defined as a behavior node, and the key behavior node is determined in each behavior node based on the model of the attention mechanism, and is the user behavior corresponding to the target interaction information.

Further, when the interaction information of the user to be transitioned in the preset interaction dimension of the target service is input to the target attribution model for processing, the target interaction information can be selected from the interaction information in a scoring mode by taking the accuracy and the efficiency of determining the target interaction information into consideration, and the method is specifically realized as follows:

determining interaction information of the user to be transitioned in at least one preset interaction dimension of the target service in the history interaction information of the user to be transitioned; inputting interaction information of at least one preset interaction dimension into an initial attribution model for prediction, and obtaining a prediction score of each piece of interaction information; ordering each piece of interaction information according to the prediction scores; and selecting target interaction information of the user to be transited from the interaction information of at least one preset interaction dimension according to the sequencing result.

Specifically, the historical interaction information refers to interaction information of a user associated store or commodity, including but not limited to activity participation information generated when the user participates in an activity, and commodity purchase information generated when the user purchases the commodity; the prediction score is a score obtained by scoring the influence degree of each piece of interaction information by the attribution model, the prediction score is used for representing the influence degree of the interaction information on the change of the user grade, and the prediction score is positively correlated with the influence degree, namely, the higher the prediction score is, the higher the influence degree is; the ranking result refers to a ranking result obtained by ranking according to the prediction score of each piece of interaction information, namely an interaction information sequence.

Based on the above, in the historical interaction information of the user to be transited, determining the interaction information of the user to be transited in at least one preset interaction dimension of the target service, inputting the interaction information of the at least one preset interaction dimension into the initial attribution model for prediction, and obtaining a prediction score of each piece of interaction information. And sequencing each piece of interaction information according to the prediction scores, and selecting target interaction information of the user to be transited from the interaction information of at least one preset interaction dimension according to the sequencing result. When selecting, a prediction score threshold value can be preset, and the interaction information which is larger than the prediction score threshold value is used as target interaction information; the selection quantity of the target interaction information can be preset, and the target interaction information with the set quantity is selected according to the order of the predictive scores from high to low.

Along the above example, 6 pieces of interaction information corresponding to one user A to be transited are input to a target attribution model to be predicted, and a score 8 of the interaction information 1, a score 1 of the interaction information 2, a score 5 of the interaction information 3, a score 2 of the interaction information 4, a score 6 of the interaction information 5, and a score 3 of the interaction information 6 are respectively obtained, the interaction information is ordered according to the order of the scores from high to low, and then the interaction information 1 with the high score is selected as the interaction information with larger grade change influence degree of the user A to be transited.

In fig. 3 (b), as shown in fig. 3 (b), an initial vector of each user is used as an input of a graph in the attribution model, an interaction information sequence is used as an input of an embedded layer, an attention score of each interaction information is obtained through processing of an attention module, and then the score of each interaction information is obtained through a pooling layer, a multi-layer perceptron and an activation function.

In summary, the interaction information is ranked based on the prediction score of each interaction information, so that the target interaction information is determined in the ranked interaction information, and the accuracy of determining the target interaction information is improved.

Further, in order to improve the prediction accuracy of the target attribution model, the initial attribution model needs to be trained, which is specifically implemented as follows:

determining training interaction information of training users in the training user set in the preset interaction dimension of the target service; determining a sample label of the training user according to the attribute information of the training user; inputting the training interaction information into an initial attribution model for training, and obtaining a prediction score of the training interaction information; training the initial attribution model based on the sample label and the predictive score until a target attribution model meeting training stopping conditions is obtained.

Specifically, the training interaction information refers to user behavior information for performing training of attribution models, including, but not limited to, information of user participation activities, commodity purchase information, and the like; training the attribute information of the user to determine whether the user has grade change; correspondingly, the sample label of the training user is whether the grade change occurs or not.

Based on the training interaction information of the training user in the target service preset interaction dimension in the training user set is determined, and a sample label of the training user is determined according to the attribute information of the training user. And inputting the training interaction information into the initial attribution model for training to obtain a prediction score of the training interaction information. Training the initial attribution model based on the sample label and the predictive score until a target attribution model meeting a training stopping condition is obtained.

According to the above example, whether the user generates the grade change is used as a sample label, and training is carried out on the initial attribution model based on the interaction information of the user, so that a trained target attribution model is obtained.

In summary, the target attribution model is obtained by training the initial attribution model, so that the accuracy of determining the target interaction information based on the target attribution model is improved.

Step S208: and creating a grade change strategy associated with the user to be transitioned based on the target interaction information.

Specifically, after the interaction information of the user to be transitioned in the preset interaction dimension of the target service is input to the target attribution model for processing, and the target interaction information of the user to be transitioned is obtained, a grade change strategy associated with the user to be transitioned can be created based on the target interaction information, wherein the grade change strategy refers to a strategy for assisting the user in grade change, and the grade change strategy can be an activity recommendation strategy or a commodity recommendation strategy and the like.

Based on the interaction information of the user to be transitioned in the preset interaction dimension of the target service is input into the target attribution model for processing, after the target interaction information of the user to be transitioned is obtained, a grade change strategy of the associated user to be transitioned is created based on the target interaction information, and then the grade change is assisted by the user.

In practical application, the target interaction information can represent interaction information with a large degree of influence on the user grade change, and a grade change strategy corresponding to the user to be transited can be generated according to the determined target interaction information, so that the grade change strategy is used for assisting the user in grade change, the grade change rate of the user is improved, the number of high-grade users in a store can be further improved, and the value of the store is improved.

Further, after determining a level change policy for each user to be transitioned, in order to facilitate the change of the user level, a level change task corresponding to the user to be transitioned needs to be created according to the level change policy, so as to assist the level change of the user, which is specifically implemented as follows:

creating a grade change task of the target service related to the user to be transitioned based on the grade change strategy; and the grade change task is issued to the user to be transitioned and is used for assisting the user grade change of the user to be transitioned.

Specifically, the grade change task is a personalized task created by a pointer for each user to be transitioned, the grade change task is matched with a grade change strategy of the user to be transitioned, the grade change task can be a commodity pushing task, an activity pushing task and the like, and the grade change task is a task created aiming at target interaction information of the user to be transitioned and is used for promoting grade change of the user.

Based on the above, after the grade change strategy of the user to be transitioned is created according to the determined target interaction information, at least one grade change task of the target service related to the user to be transitioned is created according to the grade change strategy, and then the grade change task is sent to the user to be transitioned at a set time, so that the user to be transitioned is assisted in grade change.

For example, when it is determined that the level change policy of the user a to be transitioned is activity push, a new activity is created for the user a to be transitioned according to the activity in which the user a to be transitioned participates, and the activity is pushed to the user at a fixed time point, where eight points in the evening can be selected, and typically, the probability that the user sees and has time to carefully look at this time point is high. It should be noted that the time point may be determined according to actual requirements.

In summary, in the information processing method provided in one embodiment of the present disclosure, a non-transition user group is constructed by selecting a user whose user level is unchanged from a user set; according to the service participation information of the non-transition users in the target service in the non-transition user group, screening the users to be transitioned, the user grades of which are to be changed, in the non-transition user group; inputting interaction information of a user to be transitioned in a target service preset interaction dimension into a target attribution model for processing to obtain target interaction information of the user to be transitioned; and creating a grade change strategy for associating the user to be transited based on the target interaction information. According to the business participation information of the non-transition users, the users to be transited with the user grade to be changed are screened from the non-transition user group, the accuracy of user screening can be improved, the interaction information is processed through the attribution model, the target interaction information is determined in the interaction information, the accuracy of target interaction information determination can be improved, and then a grade change strategy corresponding to the target business information is created, so that the grade change of the users to be transited is promoted, and the accurate touch of the users is realized.

The information processing method provided in the present specification will be further described with reference to fig. 4 by taking an application of the information processing method to user information processing as an example. Fig. 4 is a flowchart of a processing procedure of an information processing method according to an embodiment of the present disclosure, which specifically includes the following steps.

Step S402: users are ranked according to their total life cycle value.

In the field of electronic commerce operation, through analyzing the data of the stores, the level change of the users of the stores can be found along with the time, so that key factors influencing the level change of the users are found out, personalized intervention strategies are formulated for the users, the number of high-value users in the stores can be increased, and the value of the stores is further increased. And classifying the users according to the total life cycle value of the users, calculating the total life cycle value of the users in a period of time, and classifying the users.

Step S404: and acquiring a user group formed by the users associated with the store in preset time, and constructing a user map based on the user information of the users in the user group.

And acquiring users who have purchase behaviors in a plurality of stores within one month, taking the stores, the users, the commodities and the store activities as nodes, and taking interaction data among the nodes as sides to construct a user map. The specific interaction data comprises the collection, additional purchase and purchasing behavior of the user in the store or for the commodity, the store activities performed by the store, the user participated in the store activities, and the like

Step S406: and pre-training the nodes contained in the user map to obtain initial vectors of the nodes, and constructing a transition prediction model based on the initial vectors.

Based on the user map, the node is pre-trained by adopting GraphSage (graphics Xi Suanfa) algorithm, and an initial vector of the node is obtained. The user who had made level transition within three days of history was taken as a seed user, where the transition from the lower level to the higher level was noted as seed user 1, and the transition from the higher level to the lower level was noted as seed user 2. And respectively taking the two types of seed users as positive samples, obtaining negative samples through a negative sampling method, and constructing two transition prediction models on the basis of the pre-training vectors of the graph.

Step S408: and screening potential transition users with the grades to be changed based on the transition prediction model.

And predicting the users with hierarchical transition by adopting the two transition prediction models, wherein the users with the prediction scores larger than a certain threshold value are selected as potential transition users.

Step S410: and inputting the behavior information sequence of the potential user into the attribution model for processing to obtain key behavior information in the behavior information sequence.

The key behavior information is found in the behavior information sequence of the user by adopting a method of multi-example learning MIL (Multiple Instance Learning). Behavior information which possibly affects user level transition is determined in the behavior information of the user, for example, the user makes a single N times in a store, the GMV (commodity transaction total) of the user in the store reaches a certain threshold value, the user participates in a certain activity of the store, and the like, and the behavior information is defined as different historical behavior nodes, so that a plurality of historical behavior nodes of one user form a node sequence. And screening out key behavior nodes from the node sequence by adopting an attribution model based on an attention mechanism.

When the attribution model is trained, the label of the training data is whether the user generates level transition, and in the training process of the model, the key behavior nodes are assigned higher attention scores. In the model prediction process, the behavior node with higher attention score is used as a key node affecting user transition, namely the reason of the user transition.

Step 412: a level change policy is created for the user based on the key behavior information.

And according to the determined key nodes affecting the user transition, performing personalized intervention on the user. For example, if the key node is that the user places a single order N times in the store, the merchant can increase the user retention rate by pushing discount commodities; if the key node participates in a certain activity of the store for the user, the merchant can push a similar activity to further promote the level transition of the user.

Corresponding to the above method embodiments, the present disclosure further provides an embodiment of an information processing apparatus, and fig. 5 shows a schematic structural diagram of an information processing apparatus provided in one embodiment of the present disclosure. As shown in fig. 5, the apparatus includes:

a building module 502 configured to select a user whose user level is unchanged from the user set to build a non-transition user group;

a screening module 504, configured to screen the user to be changed in the user class from the non-transition user group according to the service participation information of the non-transition user in the target service;

the input module 506 is configured to input the interaction information of the user to be transitioned in the preset interaction dimension of the target service to a target attribution model for processing, so as to obtain target interaction information of the user to be transitioned;

a creation module 508 configured to create a level change policy associated with the user to be transitioned based on the target interaction information.

In an alternative embodiment, the building block 502 is further configured to:

In an alternative embodiment, the screening module 504 is further configured to:

inputting service participation information of non-transition users in the non-transition user group in the target service into a first target prediction model and a second target prediction model for prediction; screening a first user to be transited, the user grade of which is to be changed, from the non-transition user group according to a first prediction result corresponding to the first target prediction model, and screening a second user to be transited, the user grade of which is to be changed, from the non-transition user group according to a second prediction result corresponding to the second target prediction model;

And forming a user to be transitioned by the first user to be transitioned and the second user to be transitioned.

In an alternative embodiment, the input module 506 is further configured to:

training of the target attribution model is as follows: determining training interaction information of training users in the training user set in the preset interaction dimension of the target service; determining a sample label of the training user according to the attribute information of the training user; inputting the training interaction information into an initial attribution model for training, and obtaining a prediction score of the training interaction information; training the initial attribution model based on the sample label and the predictive score until a target attribution model meeting training stopping conditions is obtained.

In an alternative embodiment, the input module 506 is further configured to:

the determining of the business participation information comprises the following steps: reading resource participation information, operation information and/or browsing information of a non-transition user in the target service in the non-transition user group; taking the resource participation information, the operation information and/or the browsing information as service participation information; correspondingly, the determining of the interaction information comprises the following steps: reading resource interaction information, resource score information and/or task participation information of the user to be transitioned in the preset interaction dimension of the target service; and taking the resource interaction information, the resource score information and/or the task participation information as interaction information.

In an alternative embodiment, the creation module 508 is further configured to:

In summary, an embodiment of the present disclosure provides an information processing apparatus that constructs a non-transition user group by selecting a user whose user level is unchanged from a user set; according to the service participation information of the non-transition users in the target service in the non-transition user group, screening the users to be transitioned, the user grades of which are to be changed, in the non-transition user group; inputting interaction information of a user to be transitioned in a target service preset interaction dimension into a target attribution model for processing to obtain target interaction information of the user to be transitioned; and creating a grade change strategy for associating the user to be transited based on the target interaction information. According to the business participation information of the non-transition users, the users to be transited with the user grade to be changed are screened from the non-transition user group, the accuracy of user screening can be improved, the interaction information is processed through the attribution model, the target interaction information is determined in the interaction information, the accuracy of target interaction information determination can be improved, and then a grade change strategy corresponding to the target business information is created, so that the grade change of the users to be transited is promoted, and the accurate touch of the users is realized.

The above is a schematic scheme of an information processing apparatus of the present embodiment. It should be noted that, the technical solution of the information processing apparatus and the technical solution of the information processing method belong to the same concept, and details of the technical solution of the information processing apparatus, which are not described in detail, can be referred to the description of the technical solution of the information processing method.

Fig. 6 illustrates a block diagram of a computing device 600 provided in accordance with one embodiment of the present description. The components of computing device 600 include, but are not limited to, memory 610 and processor 620. The processor 620 is coupled to the memory 610 via a bus 630 and a database 650 is used to hold data.

Computing device 600 also includes access device 640, access device 640 enabling computing device 600 to communicate via one or more networks 660. Examples of such networks include public switched telephone networks (PSTN, public Switched Telephone Network), local area networks (LAN, local Area Network), wide area networks (WAN, wide Area Network), personal area networks (PAN, personal Area Network), or combinations of communication networks such as the internet. The access device 640 may include one or more of any type of network interface, wired or wireless, such as a network interface card (NIC, network interface controller), such as an IEEE802.11 wireless local area network (WLAN, wireless Local Area Network) wireless interface, a worldwide interoperability for microwave access (Wi-MAX, worldwide Interoperability for Microwave Access) interface, an ethernet interface, a universal serial bus (USB, universal Serial Bus) interface, a cellular network interface, a bluetooth interface, a near field communication (NFC, near Field Communication) interface, and so forth.

In one embodiment of the present application, the above-described components of computing device 600, as well as other components not shown in FIG. 6, may also be connected to each other, such as by a bus. It should be understood that the block diagram of the computing device illustrated in FIG. 6 is for exemplary purposes only and is not intended to limit the scope of the present application. Those skilled in the art may add or replace other components as desired.

Computing device 600 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., tablet, personal digital assistant, laptop, notebook, netbook, etc.), mobile phone (e.g., smart phone), wearable computing device (e.g., smart watch, smart glasses, etc.), or other type of mobile device, or a stationary computing device such as a desktop computer or personal computer (PC, personal Computer). Computing device 600 may also be a mobile or stationary server. Wherein the processor 620 is configured to execute computer-executable instructions that, when executed by the processor, perform the steps of the information processing method described above.

The foregoing is a schematic illustration of a computing device of this embodiment. It should be noted that, the technical solution of the computing device and the technical solution of the information processing method belong to the same concept, and details of the technical solution of the computing device, which are not described in detail, can be referred to the description of the technical solution of the information processing method.

An embodiment of the present disclosure also provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the above-described information processing method.

The above is an exemplary version of a computer-readable storage medium of the present embodiment. It should be noted that, the technical solution of the storage medium and the technical solution of the information processing method belong to the same concept, and details of the technical solution of the storage medium which are not described in detail can be referred to the description of the technical solution of the information processing method.

An embodiment of the present specification also provides a computer program, wherein the computer program, when executed in a computer, causes the computer to execute the steps of the above information processing method.

The above is an exemplary version of a computer program of the present embodiment. It should be noted that, the technical solution of the computer program and the technical solution of the information processing method belong to the same conception, and details of the technical solution of the computer program, which are not described in detail, can be referred to the description of the technical solution of the information processing method.

The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

The computer instructions include computer program code that may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the computer readable medium contains content that can be appropriately scaled according to the requirements of jurisdictions in which such content is subject to legislation and patent practice, such as in certain jurisdictions in which such content is subject to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of combinations of actions, but it should be understood by those skilled in the art that the embodiments are not limited by the order of actions described, as some steps may be performed in other order or simultaneously according to the embodiments of the present disclosure. Further, those skilled in the art will appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily all required for the embodiments described in the specification.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.

The preferred embodiments of the present specification disclosed above are merely used to help clarify the present specification. Alternative embodiments are not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the teaching of the embodiments. The embodiments were chosen and described in order to best explain the principles of the embodiments and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. This specification is to be limited only by the claims and the full scope and equivalents thereof.

Claims

1. An information processing method, comprising:

2. The method of claim 1, the selecting a user of unchanged user level in the set of users to construct a non-transition user group, comprising:

reading historical transition information of users in the user set in a preset time interval;

selecting non-transition users with unchanged user grades in the user set based on the historical transition information;

and constructing a non-transition user group based on the non-transition user.

3. The method according to claim 1, wherein the selecting, according to the service participation information of the non-transition users in the non-transition user group in the target service, the user to be transitioned whose user level is to be changed in the non-transition user group includes:

reading service participation information of non-transition users in the non-transition user group in the target service, and inputting the service participation information into a target prediction model for prediction;

and screening the users to be changed in the user level in the non-transition user group according to the prediction result.

4. A method according to claim 3, the training of the target prediction model being as follows:

Selecting an upward transition user from a training user set, and taking transition participation information of the upward transition user as a first positive sample;

selecting a downward transition user from the training user set, taking transition participation information of the downward transition user as a second positive sub-sample, and forming a positive sample by the first positive sub-sample and the second positive sub-sample;

performing negative sampling on the training user set to obtain a negative sample;

and training the initial prediction model based on the positive sample and the negative sample until a target prediction model meeting the training stop condition is obtained.

5. The method of claim 4, the training an initial predictive model based on the positive and negative samples, until a target predictive model step is performed that meets a training stop condition, further comprising:

constructing a user map based on training user information of training users in the training user set;

constructing an initial vector of a user node in the user map;

a first initial prediction model and a second initial prediction model are generated based on the initial vector of each training user, and the first initial prediction model and the second initial prediction model are used as initial prediction models.

6. The method of claim 5, the training an initial predictive model based on the positive and negative samples until a target predictive model is obtained that meets a training stop condition, comprising:

training the first initial prediction model based on the first positive sub-sample and the negative sub-sample until a first target prediction model meeting a training stop condition is obtained;

training the second initial prediction model based on the second positive sub-sample and the negative sub-sample until a second target prediction model meeting a training stop condition is obtained;

and taking the first target prediction model and the second target prediction model as target prediction models.

7. The method according to claim 6, wherein inputting the service participation information of the non-transition users in the non-transition user group in the target service into a target prediction model for prediction, and selecting the user to be transitioned with the user level to be changed from the non-transition user group according to the prediction result, comprises:

inputting service participation information of non-transition users in the non-transition user group in the target service into a first target prediction model and a second target prediction model for prediction;

Screening a first user to be transited, the user grade of which is to be changed, from the non-transition user group according to a first prediction result corresponding to the first target prediction model, and screening a second user to be transited, the user grade of which is to be changed, from the non-transition user group according to a second prediction result corresponding to the second target prediction model;

8. The method of claim 1, wherein inputting the interaction information of the user to be transitioned in the target service preset interaction dimension into a target attribution model for processing to obtain the target interaction information of the user to be transitioned, comprises:

determining interaction information of the user to be transitioned in at least one preset interaction dimension of the target service in the history interaction information of the user to be transitioned;

inputting interaction information of at least one preset interaction dimension into an initial attribution model for prediction, and obtaining a prediction score of each piece of interaction information;

ordering each piece of interaction information according to the prediction scores;

and selecting target interaction information of the user to be transited from the interaction information of at least one preset interaction dimension according to the sequencing result.

9. The method of claim 4, the training of the target attribution model being as follows:

determining training interaction information of training users in the training user set in the preset interaction dimension of the target service;

determining a sample label of the training user according to the attribute information of the training user;

inputting the training interaction information into an initial attribution model for training, and obtaining a prediction score of the training interaction information;

training the initial attribution model based on the sample label and the predictive score until a target attribution model meeting training stopping conditions is obtained.

10. The method of claim 1, the determining of the business participation information comprising:

reading resource participation information, operation information and/or browsing information of a non-transition user in the target service in the non-transition user group;

taking the resource participation information, the operation information and/or the browsing information as service participation information;

correspondingly, the determining of the interaction information comprises the following steps:

reading resource interaction information, resource score information and/or task participation information of the user to be transitioned in the preset interaction dimension of the target service;

And taking the resource interaction information, the resource score information and/or the task participation information as interaction information.

11. The method of claim 1, further comprising, after the step of creating a level change policy associated with the user to be transitioned based on the target interaction information is performed:

creating a grade change task of the target service related to the user to be transitioned based on the grade change strategy;

and the grade change task is issued to the user to be transitioned and is used for assisting the user grade change of the user to be transitioned.

12. An information processing apparatus comprising:

13. A computing device, comprising:

a memory and a processor;

the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, which when executed by the processor, implement the steps of the information processing method according to any one of claims 1 to 11.

14. A computer-readable storage medium storing computer-executable instructions which, when executed by a processor, implement the steps of the information processing method of any one of claims 1 to 11.