CN111681049B

CN111681049B - Processing method of user behavior, storage medium and related equipment

Info

Publication number: CN111681049B
Application number: CN202010502128.3A
Authority: CN
Inventors: 黄昕虹
Original assignee: Guangzhou Shiyuan Electronics Thecnology Co Ltd
Current assignee: Guangzhou Shiyuan Electronics Thecnology Co Ltd
Priority date: 2020-06-04
Filing date: 2020-06-04
Publication date: 2023-08-11
Anticipated expiration: 2040-06-04
Also published as: CN111681049A

Abstract

The embodiment of the application discloses a processing method of user behaviors, a storage medium and related equipment, wherein the method comprises the following steps: acquiring behavior data of a plurality of objects; constructing a directed graph based on behavior data of a plurality of objects, wherein the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges; searching the directed graph, determining at least one behavior path for transferring non-target behaviors to target behaviors, and a weight value of each behavior path, wherein the number of different behaviors contained in the behavior path is smaller than or equal to a preset number. Therefore, the embodiment of the application can achieve the technical effects of improving the prediction accuracy, improving the user experience and improving the user viscosity, and further solves the technical problems of low processing accuracy and poor user experience of the user behavior in the related technology.

Description

Processing method of user behavior, storage medium and related equipment

Technical Field

The present application relates to the field of internet, and in particular, to a method for processing user behavior, a storage medium, and related devices.

Background

Currently, user behavior analysis and behavior prediction generally predict whether a user will perform a specific behavior by extracting features of the user, and use a classification model, for example, in an e-commerce shopping scenario, the purchasing potential of a user-commodity pair can be captured by extracting features, which is specifically implemented as follows: applying an extraction application gradient lifting decision tree (Gradient Boosting Decision Tree, GBDT) as a training model; constructing a user prediction model by using mathematical and statistical models and combining theories such as user purchasing behavior, process, influence factors and the like; constructing a selection model based on potential factors based on the target behavior sequence data, and further predicting purchase decisions of users in a purchase period; mathematical models such as linear regression and logarithmic models are used to predict the probability of multiple purchases by a customer over a period of time.

However, in an actual scene, since the iteration speed of the APP function is high, the user behavior habit change frequency is high, the difficulty of user behavior prediction is increased, and whether a user performs specific behaviors or not can be predicted only based on single user feature modeling, so that the user behavior prediction accuracy is low and the user experience is poor.

In view of the above problems, no effective solution has been proposed at present.

Disclosure of Invention

The embodiment of the application provides a processing method, a storage medium and related equipment for user behaviors, which at least solve the technical problems of low accuracy of user behavior prediction and poor user experience in the related technology.

According to an aspect of the embodiment of the present application, there is provided a method for processing user behavior, including: acquiring behavior data of a plurality of objects; constructing a directed graph based on behavior data of a plurality of objects, wherein the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges; searching the directed graph, determining at least one behavior path for transferring non-target behaviors to target behaviors, and a weight value of each behavior path, wherein the number of different behaviors contained in the behavior path is smaller than or equal to a preset number.

Optionally, building the directed graph based on behavior data of the plurality of objects includes: dividing behavior data of a plurality of objects to obtain a plurality of behavior sequences; constructing a training set and a testing set based on a plurality of behavior sequences, wherein the training set comprises: a positive sample training set and a negative sample training set, the test set comprising: a positive sample test set and a negative sample test set, wherein the positive sample is a behavior sequence containing target behaviors, and the negative sample is a behavior sequence not containing target behaviors; constructing a first directed graph by using the training set; testing the first directed graph based on the test set to obtain a test result; and under the condition that the test result does not meet the preset condition, updating the first directed graph until the test result meets the preset condition, and obtaining the directed graph.

Optionally, constructing the initial directed graph using the training set includes: constructing a second directed graph by using the positive sample training set; determining a first behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the positive sample edge-shearing rate; determining a second behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the negative-sample edge-shearing rate; and cutting edges of the second directed graph based on the first behavior sequence and the second behavior sequence to obtain a first directed graph.

Optionally, updating the first directed graph includes: updating the positive sample edge shearing rate and the negative sample edge shearing rate based on a preset step length to obtain updated positive sample edge shearing rate and negative sample edge shearing rate; determining a third behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the updated positive sample edge-shearing rate; determining a fourth behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the updated negative-sample edge-shearing rate; and cutting edges of the second directed graph based on the third behavior sequence and the fourth behavior sequence to obtain a third directed graph, wherein the third directed graph is determined to be the directed graph under the condition that a test result corresponding to the second directed graph meets a preset condition.

Optionally, constructing the second directed graph using the positive sample training set includes: processing the positive sample training set to determine nodes and edges in the second directed graph; extracting characteristics of behavior data of two behaviors connected by edges to obtain characteristic data corresponding to the edges; and obtaining the weight value of the edge based on the characteristic data corresponding to the edge.

Optionally, constructing the training set and the testing set based on the plurality of behavior sequences includes: determining a positive sample data set and a negative sample data set based on the plurality of behavior sequences; randomly dividing a positive sample data set to obtain a positive sample training set and a positive sample testing set; and randomly dividing the negative sample data set to obtain a negative sample training set and a negative sample testing set.

Optionally, determining whether the test result meets the preset condition includes: determining a first recognition rate based on a first test result of the positive sample test set, wherein the first recognition rate is used for representing the duty ratio of the positive sample test result in the first test result; determining a second recognition rate based on a second test result of the negative sample test set, wherein the second recognition rate is used for representing the duty ratio of the positive sample test result in the second test result; after updating the positive sample trimming rate and the negative sample trimming rate, acquiring a first variation of the first recognition rate and a second variation of the second recognition rate; judging whether the first recognition rate is smaller than a preset value, whether the first recognition rate is smaller than the second recognition rate, and whether the first variation is smaller than the second variation; if the first recognition rate is smaller than the preset value, the first recognition rate is smaller than the second recognition rate, and the first variation is smaller than the second variation, determining that the test result meets the preset condition; if the first recognition rate is greater than or equal to a preset value, the first recognition rate is greater than or equal to the second recognition rate, or the first variation is greater than or equal to the second variation, determining that the test result does not meet the preset condition.

Optionally, before dividing the behavior data of the plurality of objects to obtain a plurality of behavior sequences, the method further includes: matching the behavior data with preset behavior data; if the behavior data is successfully matched with the preset behavior data, the behavior data is removed, and the cleaned behavior data is obtained; and dividing the washed behavior data to obtain a plurality of behavior sequences.

Optionally, before constructing the training set and the test set based on the plurality of behavior sequences, the method further comprises: judging whether two adjacent behavior sequences are identical; if the two adjacent behavior sequences are the same, any one of the two adjacent behavior sequences is eliminated to obtain a cleaned behavior sequence; and constructing a training set and a testing set based on the washed behavior sequence.

Optionally, after matching the target behavior sequence with the at least one behavior path and determining the target behavior path, the method further includes: acquiring a target behavior sequence of a target object; matching the target behavior sequence with at least one behavior path to determine a target behavior path; acquiring a weight value of a target behavior path; and determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior.

Optionally, after determining the push target information, the method further includes: monitoring whether a target object executes target behaviors; and updating the directed graph based on the target behavior sequence under the condition that the target object does not execute the target behavior.

Optionally, before matching the target behavior sequence with the at least one behavior path and determining the target behavior path, the method further includes: matching the target behavior data of the target behavior sequence with preset behavior data; if the target behavior data is successfully matched with the preset behavior data, the target behavior data is removed, and a cleaned target behavior sequence is obtained; and matching the cleaned target behavior sequence with at least one behavior path to determine a target behavior path.

According to an aspect of the embodiment of the present application, there is provided a method for processing user behavior, including: acquiring a target behavior sequence of a target object; matching the target behavior sequence with at least one behavior path to determine a target behavior path, wherein the at least one behavior path is a behavior path for transferring non-target behaviors determined by searching a directed graph to the target behaviors, the number of different behaviors contained in the behavior path is smaller than or equal to a preset number, the directed graph is constructed based on behavior data of a plurality of objects, and the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges; acquiring a weight value of a target behavior path, wherein the weight value is determined by searching a directed graph; and determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior.

According to another aspect of the embodiments of the present application, there is also provided a computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the above-described method steps.

According to another aspect of the embodiment of the present application, there is also provided an electronic device, including: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the method steps described above.

In the embodiment of the application, the directed graph is constructed based on the behavior data of a plurality of objects, the directed graph is searched, at least one behavior path for transferring the non-target behavior to the target behavior is determined, and the target behavior path can be determined by matching the target behavior sequence of the target object with the at least one behavior path, so that the aim of constructing the weighted directed graph by utilizing the historical behavior data of the whole user group and predicting the future behavior path of the user through the weighted directed graph is fulfilled, the technical effects of improving the prediction accuracy, improving the user experience feeling and improving the user viscosity are achieved, and the technical problems of low accuracy and poor user experience feeling of the user behavior prediction in the related art are solved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:

FIG. 1 is a flow chart of a first method of processing user behavior according to an embodiment of the application;

FIG. 2 is a flow chart of a second method of processing user behavior according to an embodiment of the application;

FIG. 3 is a flow chart of a third method of processing user behavior according to an embodiment of the present application;

FIG. 4 is a flow chart of a fourth method of processing user behavior according to an embodiment of the application;

FIG. 5 is a flowchart of an alternative engine side according to an embodiment of the present application;

FIG. 6 is a flow chart of an alternative engine side construction dataset according to an embodiment of the present application;

FIG. 7 is a flow chart of an alternative server side according to an embodiment of the application;

FIG. 8 is a schematic diagram of a processing device for user behavior according to an embodiment of the application;

FIG. 9 is a schematic diagram of a processing device for another user action according to an embodiment of the application; and

fig. 10 is a schematic diagram of an electronic device according to an embodiment of the application.

Detailed Description

In order that those skilled in the art will better understand the present application, a technical solution in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present application without making any inventive effort, shall fall within the scope of the present application.

When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the application as detailed in the accompanying claims.

It should be noted that the terms "first," "second," and the like in the description and the claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Furthermore, in the description of the present application, unless otherwise indicated, "a plurality" means two or more. "and/or", describes an association relationship of an association object, and indicates that there may be three relationships, for example, a and/or B, and may indicate: a exists alone, A and B exist together, and B exists alone. The character "/" generally indicates that the context-dependent object is an "or" relationship.

Example 1

The embodiment of the application provides a processing method of household behaviors, which is applied to electronic equipment, such as a computer terminal, a smart phone, a tablet personal computer and the like, but is not limited to the method.

Fig. 1 is a flowchart of a first method for processing user behavior according to an embodiment of the present application, as shown in fig. 1, the method includes the steps of:

step S102, behavior data of a plurality of objects are obtained;

the object in the above steps may be a user using an APP installed on the electronic device, which APP may be an internet APP, for example, but not limited to, a class optimization master APP.

Different users usually perform different behavior operations when using the same APP, so that in order to accurately predict the user behavior of each user, historical behavior data of all users in the APP may be obtained, for example, behavior data of all users for one week may be obtained. In order to obtain historical behavior data of the user, the behavior data of the user can be recorded in real time in the process of using the APP by the user and stored in a specific database based on the operation time, so that the behavior data of a specific time period can be queried from the specific database according to the requirement.

Step S104, constructing a directed graph based on behavior data of a plurality of objects, wherein the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges;

in this embodiment, in order to predict a future behavior path of a user, a weighted directed graph may be constructed by using historical behavior data of the user, where different behaviors of the user may be represented by different nodes, and a transition relationship between behaviors corresponding to the nodes may be represented by edges directly connected to the nodes.

Further, by using characteristics such as the number of users of the behavior pairs, the occurrence frequency of the behavior pairs, the number of users of the behavior nodes transferred to other nodes, the occurrence frequency of the behavior nodes, and the like, the weight value of the edge is calculated and used for representing the importance degree of the transfer relationship between the nodes, wherein the more important the transfer between the nodes, the larger the weight value of the corresponding edge is, which indicates that more users have the behavior.

Step S106, searching the directed graph, determining at least one behavior path for transferring the non-target behavior to the target behavior, and a weight value of each behavior path, wherein the number of different behaviors contained in the behavior path is smaller than or equal to the preset number.

The target behavior in the above step may be a predefined key behavior, for example, a function that is used in the APP in a small amount, but the operator wants the user to touch, and the non-target behavior may refer to other behaviors than the key behavior, for example, taking the class optimization master APP as an example, and the target behavior may be a function of using a release job in the broadcasting station, but is not limited thereto. The preset number can be determined according to the prediction needs of the user behaviors in different application scenes, and the application is not particularly limited to the prediction needs.

In an exemplary embodiment of the present application, a directed graph may be searched to obtain paths from all non-target behaviors to target behaviors within a certain step number, and the weight value of the path may be obtained by obtaining the sum of the weight values of all edges in the paths.

In the embodiment of the application, the directed graph is constructed based on the behavior data of a plurality of objects, and the directed graph is searched, and at least one behavior path from the non-target behavior to the target behavior is determined, so that the aim of constructing the weighted directed graph by utilizing the historical behavior data of the whole user group is fulfilled, the technical effects of improving the prediction accuracy, improving the user experience feeling and improving the user viscosity are achieved, and the technical problems of low accuracy and poor user experience feeling of the user behavior prediction in the related technology are solved.

Example 2

As shown in fig. 2, the method comprises the steps of:

step S201, behavior data of a plurality of objects are obtained;

step S202, matching the behavior data with preset behavior data;

the preset behavior data in the above steps may be predefined noise behaviors and stains, for example, class optimization master APP, and the noise behaviors may include, but are not limited to: modifying nicknames, setting avatars, and the like.

Step S203, if the behavior data is successfully matched with the preset behavior data, the behavior data is removed, and the cleaned behavior data is obtained;

in an exemplary embodiment of the present invention, after the historical behavior data of the total number of users is obtained, the obtained historical behavior data may be cleaned based on the preset behavior data, so as to obtain clean behavior data.

Step S204, dividing the cleaned behavior data to obtain a plurality of behavior sequences;

in an exemplary embodiment of the present invention, a rule for dividing behavior data may be determined based on factors such as APP usage scenario and user behavior habit, and historical behavior data of the same user may be divided according to the rule to obtain a behavior sequence. For example, the rule may be that the user does not operate for 10 minutes, and then it is determined that the user starts a new behavior sequence.

Step S205, judging whether two adjacent behavior sequences are the same;

in an exemplary embodiment of the present invention, since the user often performs repeated operations during the APP usage process, after the behavior data is divided, a repeated behavior sequence may be obtained, and therefore, it is necessary to clean adjacent repeated behavior sequences.

It should be noted that, the repeated clicking actions of the user can be cleaned, so as to solve the problem that the action log is repeatedly reported.

Step S206, if two adjacent behavior sequences are the same, any one of the two adjacent behavior sequences is eliminated, and a cleaned behavior sequence is obtained;

in an exemplary embodiment of the present invention, for adjacent repetitive behavior sequences, only one of the behavior sequences may be retained, and other repetitive behavior sequences may be eliminated, resulting in a clean behavior sequence.

Step S207, a training set and a testing set are constructed based on the cleaned behavior sequence, wherein the training set comprises: a positive sample training set and a negative sample training set, the test set comprising: a positive sample test set and a negative sample test set, wherein the positive sample is a behavior sequence containing target behaviors, and the negative sample is a behavior sequence not containing target behaviors;

The training set in the above step is used for constructing the directed graph, the testing set is used for testing the directed graph, determining whether the constructed directed graph meets the prediction requirement, and updating the directed graph under the condition that the constructed directed graph does not meet the prediction requirement. Positive samples may refer to behavior sequences that contain critical behavior and negative samples may refer to behavior sequences that do not contain critical behavior.

In the above embodiment of the present invention, the training set and the test set may be constructed as follows: determining a positive sample data set and a negative sample data set based on the plurality of behavior sequences; randomly dividing a positive sample data set to obtain a positive sample training set and a positive sample testing set; and randomly dividing the negative sample data set to obtain a negative sample training set and a negative sample testing set.

In an exemplary embodiment of the present invention, the behavior sequence may be classified according to whether the clean behavior sequence includes the target behavior, where the behavior sequence including the target behavior is classified into the positive sample data set and vice versa. Further randomly selecting 70% of the positive sample data set as positive sample training joints, the remaining 30% as positive sample test sets, and similarly randomly selecting 70% of the negative sample data set as negative sample training joints, and the remaining 30% as negative sample test sets.

It should be noted that, for positive and negative sample training sets, the training set may be divided into a training set and a verification set based on random selection, where the training set is used to construct a directed graph, and the verification set is used to verify the effect of the directed graph in the training process, similar to the effect of the test set.

Step S208, constructing a first directed graph by using the training set;

in the above embodiment of the present invention, the first directed graph may be constructed as follows: constructing a second directed graph by using the positive sample training set; determining a first behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the positive sample edge-shearing rate; determining a second behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the negative-sample edge-shearing rate; and cutting edges of the second directed graph based on the first behavior sequence and the second behavior sequence to obtain a first directed graph.

Optionally, constructing the second directed graph by using the positive sample training set specifically includes the following steps: processing the positive sample training set to determine nodes and edges in the second directed graph; extracting characteristics of behavior data of two behaviors connected by edges to obtain characteristic data corresponding to the edges; and obtaining the weight value of the edge based on the characteristic data corresponding to the edge.

In an exemplary embodiment of the present invention, a weighted directed graph is constructed with user behavior as nodes and transitions between behaviors as edges, where there are edges between two behavior nodes, indicating that there is a user performed between the two behaviors. The characteristics of the number of users of the behavior pairs, the occurrence frequency of the behavior pairs, the number of users of the behavior nodes transferred to other nodes, the occurrence frequency of the behavior nodes and the like of the edge connection can be further extracted according to actual use requirements, and the weight value of each edge is calculated based on the characteristics.

In order to reduce connectivity of the user behavior graph, avoiding the situation that the user can reach the key behavior at any behavior point, the structured behavior graph needs to be edge-cut, and edges which are not important in the positive sample and are important in the negative sample need to be cut.

The positive sample trimming rate in the above step is used to represent the ratio of edges to be trimmed from the positive samples in all positive samples, and the negative sample trimming rate is used to represent the ratio of edges to be trimmed from the negative samples in all negative samples. It should be noted that, for the positive sample, the insignificant edges, that is, the edges with smaller weight values, need to be cut off; for negative samples, important edges need to be cut, i.e. edges with larger weight values are cut.

In an exemplary embodiment of the present invention, the beat from large to small may be performed based on the weight value of each edge, the behavior sequence of the last few ordered bits in the positive sample is selected to obtain the first behavior sequence, and the behavior sequence of the first few ordered bits in the negative sample is selected to obtain the second behavior sequence, where the number of the first behavior sequences is determined by the product of the edge trimming rate of the positive sample and the behavior sequence in the positive sample, and the number of the second behavior sequences is determined by the product of the edge trimming rate of the positive sample and the behavior sequence in the negative sample. After the first behavior sequence and the second behavior sequence which need to be cut off are determined, the corresponding edges in the directed graph constructed through the positive sample training set can be deleted, and the purpose of edge cutting of the directed graph is achieved.

Step S209, testing the first directed graph based on the test set to obtain a test result;

the test result in the step is used for representing whether the obtained prediction result is a Positive sample or a negative sample after the Positive sample test set or the negative sample test set is input to the first directed graph, wherein for the Positive sample test set, if the prediction result is the Positive sample, the test result can be determined to be a True example (True Positive); for a negative sample test set, if the predicted result is a Positive sample, then the test result may be determined to be a False Positive (False Positive)

Step S210, under the condition that the test result does not meet the preset condition, updating the first directed graph until the test result meets the preset condition, and obtaining the directed graph;

the preset condition in the above step may refer to a stopping condition of searching the optimal parameter by the directed graph, and in this embodiment, a True Positive (True Positive) and a False Positive (False Positive) are taken as result evaluation criteria.

In the above embodiment of the present invention, whether the detection result satisfies the preset condition may be determined by: determining a first recognition rate based on a first test result of the positive sample test set, wherein the first recognition rate is used for representing the duty ratio of the positive sample test result in the first test result; determining a second recognition rate based on a second test result of the negative sample test set, wherein the second recognition rate is used for representing the duty ratio of the positive sample test result in the second test result; after updating the positive sample trimming rate and the negative sample trimming rate, acquiring a first variation of the first recognition rate and a second variation of the second recognition rate; judging whether the first recognition rate is smaller than a preset value, whether the first recognition rate is smaller than the second recognition rate, and whether the first variation is smaller than the second variation; if the first recognition rate is smaller than the preset value, the first recognition rate is smaller than the second recognition rate, and the first variation is smaller than the second variation, determining that the test result meets the preset condition; if the first recognition rate is greater than or equal to a preset value, the first recognition rate is greater than or equal to the second recognition rate, or the first variation is greater than or equal to the second variation, determining that the test result does not meet the preset condition.

The first recognition rate may be an accuracy rate, the second recognition rate may be a false recognition rate, and the preset value may be an accuracy rate threshold set in advance according to actual needs, and in this embodiment, 73% is taken as an example for illustration.

In an exemplary embodiment of the present invention, the preset condition includes three stop conditions, the first stop condition being an accuracy rate of less than 73%; the second stopping condition is that the accuracy rate is reduced more than the false recognition rate after the parameters are updated, namely, the change amount of the accuracy rate after the parameters are updated is larger than the change amount of the false recognition rate; the third stop condition is that the accuracy is lower than the false recognition rate.

Optionally, updating the first directed graph specifically includes the following steps: updating the positive sample edge shearing rate and the negative sample edge shearing rate based on a preset step length to obtain updated positive sample edge shearing rate and negative sample edge shearing rate; determining a third behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the updated positive sample edge-shearing rate; determining a fourth behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the updated negative-sample edge-shearing rate; and cutting edges of the second directed graph based on the third behavior sequence and the fourth behavior sequence to obtain a third directed graph, wherein the third directed graph is determined to be the directed graph under the condition that a test result corresponding to the second directed graph meets a preset condition.

The preset step length in the above steps can be set according to actual needs, and the present application is not limited in particular.

In an exemplary embodiment of the present application, the built directed graph may be tested based on the test set, if it is determined that the test result does not meet the three stopping conditions, that is, does not meet any one of the stopping conditions, the parameters are updated according to a preset step length, and the edge of the built original directed graph (that is, the second directed graph) is cut until the three stopping conditions are met, and the final directed graph (that is, the third directed graph) is output for prediction.

By trimming the directed graph, the false recognition rate can be reduced to the greatest extent while ensuring a certain prediction accuracy rate, so that the directed graph with the best prediction effect is obtained.

Step S211, retrieving the directed graph, determining at least one behavior path for transferring the non-target behavior to the target behavior, and a weight value of each behavior path.

In the embodiment of the application, the weighted directed graph is constructed by utilizing the historical behavior data of the whole user group, so that the technical effects of improving the prediction accuracy, improving the user experience and improving the user viscosity are achieved, and the technical problems of low accuracy of user behavior prediction and poor user experience in the related technology are solved.

Example 3

As shown in fig. 3, the method comprises the steps of:

step S301, behavior data of a plurality of objects are obtained;

step S302, constructing a directed graph based on behavior data of a plurality of objects, wherein the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges;

step S303, retrieving the directed graph, determining at least one behavior path for transferring non-target behaviors to target behaviors, and a weight value of each behavior path, wherein the number of different behaviors contained in the behavior path is smaller than or equal to a preset number;

step S304, a target behavior sequence of a target object is obtained;

the target object in the above step may be a user currently using APP, or may be a key user that the operator wishes to pay attention to, and the target behavior sequence may be a series of operation behaviors currently performed by the user, where the sequence includes a plurality of behaviors performed by the user and a transition relationship between different behaviors, so that the sequence may be used as a behavior path.

Step S305, matching the target behavior data of the target behavior sequence with preset behavior data;

Step S306, if the target behavior data is successfully matched with the preset behavior data, the target behavior data is removed, and a cleaned target behavior sequence is obtained;

in an exemplary embodiment of the present invention, similar to the process of constructing the directed graph, after the target behavior sequence is obtained, the obtained target behavior sequence may be cleaned based on preset behavior data to obtain a clean behavior sequence.

Step S307, matching the cleaned target behavior sequence with at least one behavior path to determine a target behavior path;

the target behavior path in the above step may be a behavior path in which the user is predicted to most possibly reach the target behavior based on the target behavior sequence.

In an exemplary embodiment of the present invention, in order to implement prediction of a user behavior, a last behavior in a target behavior sequence may be matched with all behavior paths, and whether the last behavior exists in the behavior paths is determined, if the last behavior exists in a certain behavior path, it is determined that the target behavior sequence is successfully matched with the behavior path, so that it is determined that the behavior path is a target behavior path; if the target behavior sequence does not exist in a certain behavior path, determining that the target behavior sequence fails to match the behavior path.

It should be noted that, more than one behavior path is often successfully matched with the target behavior sequence, and one or more behavior paths with extremely large weight values can be selected as the target behavior paths further based on the weight values of the behavior paths.

Step S308, obtaining a weight value of a target behavior path;

step S309, determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior;

the target information in the above steps may be an intervention suggestion for intervention on the user behavior, and by intervention on the user needing intervention, the user may execute the target behavior in the operation process of the subsequent APP.

In an exemplary embodiment of the present invention, for a behavior path with a lower weight value, the probability that the user touches the target behavior is smaller, so that intervention is not suggested, and it is determined that the target information is not pushed; for the behavior paths with higher weight values, the user is more likely to autonomously execute the target behavior in the actual use process, so that intervention is not suggested, and target information is determined not to be pushed; in addition to the above two cases, the user has an opportunity to contact the target behavior, and to ensure that the user can contact the target behavior, the behavior of the user can be interfered, so that the target information is determined to be pushed.

Optionally, in order to accurately determine whether to push the target information based on the weight value of the target behavior path, after searching at least one behavior path from the constructed directed graph and calculating the corresponding weight value, the behavior paths may be sorted in descending order according to the weight value, and divided into behavior paths with high, medium and low level probabilities according to the ratio of 3:5:2. After the weight value of the target behavior path is obtained, the corresponding level probability can be determined, if the level probability is medium level probability, the intervention is determined to be needed, and the target information can be pushed; if the probability is high or low, no intervention is recommended and the target information may not be pushed.

It should be noted that, the dividing criteria of the three level probabilities may be set according to actual needs, and in the embodiment of the present invention, a ratio of 3:5:2 is taken as an example for illustration.

Step S310, after determining to push the target information, monitoring whether the target object executes the target behavior;

in an exemplary embodiment of the present invention, after the user behavior is interfered, it is required to monitor whether the user touches the target behavior after being interfered, and update the directed graph according to the monitoring result, so as to further improve the accuracy of the user behavior prediction.

It should be noted that if the user touches the target behavior, that is, the obtained target behavior sequence includes the target behavior, the target behavior sequence of the user at this time may be used as the test set, and the directed graph may be updated.

In step S311, when the target object does not execute the target behavior, the directed graph is updated based on the target behavior sequence.

In an exemplary embodiment of the present application, after the intervention of the user behavior, it is required to monitor whether the user touches the target behavior after the intervention, and if the user does not touch the target behavior, that is, the obtained target behavior sequence does not include the target behavior, the prediction error sequence is collected, and the directed graph is periodically analyzed and updated.

In the embodiment of the application, the weighted directed graph is constructed by utilizing the historical behavior data of the whole user group, and the future behavior path of the user is predicted by the weighted directed graph, so that the technical effects of improving the prediction accuracy, improving the user experience and improving the user viscosity are achieved, and the technical problems of low accuracy and poor user experience of the user behavior prediction in the related technology are solved.

Example 4

The embodiment of the application also provides a processing method of the household behavior, which is applied to electronic equipment such as a computer terminal, a smart phone, a tablet personal computer and the like, but is not limited to the method, and the application is not limited to the method.

Fig. 4 is a flowchart of a processing method of a fourth user behavior according to an embodiment of the present application, as shown in fig. 4, the method includes the steps of:

step S402, a target behavior sequence of a target object is obtained;

Step S404, matching the target behavior sequence with at least one behavior path to determine a target behavior path, wherein the at least one behavior path is a behavior path for transferring non-target behaviors determined by searching a directed graph to the target behaviors, the number of different behaviors contained in the behavior path is less than or equal to a preset number, the directed graph is constructed based on behavior data of a plurality of objects, and the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges;

The target behavior in the above step may be a predefined key behavior, for example, a function that is used in the APP in a small amount, but the operator wants the user to touch, and the non-target behavior may refer to other behaviors than the key behavior, for example, taking the class optimization master APP as an example, and the target behavior may be a function of using a release job in the broadcasting station, but is not limited thereto. The preset number can be determined according to the prediction needs of the user behaviors in different application scenes, and the invention is not particularly limited to the prediction needs.

In an exemplary embodiment of the present invention, a directed graph may be searched to obtain paths from all non-target behaviors to target behaviors within a certain step number, and the weight value of the path may be obtained by obtaining the sum of the weight values of all edges in the paths.

Step S406, obtaining a weight value of the target behavior path, wherein the weight value is determined by searching the directed graph;

Step S408, determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior.

In an exemplary embodiment of the present application, in order to implement prediction of a user behavior, a last behavior in a target behavior sequence may be matched with all behavior paths, and whether the last behavior exists in the behavior paths is determined, if the last behavior exists in a certain behavior path, it is determined that the target behavior sequence is successfully matched with the behavior path, so that it is determined that the behavior path is a target behavior path; if the target behavior sequence does not exist in a certain behavior path, determining that the target behavior sequence fails to match the behavior path.

Example 5

The processing method of the user behavior provided by the embodiment of the invention can be applied to the APP of the electronic equipment. In the embodiment of the present invention, the present invention will be described with reference to fig. 5 to 7 by taking class optimization master APP as an example.

In class optimization master APP, operation intervention can be performed on a specific user generating a certain real-time behavior sequence according to the actual business flow of class optimization master, and the specific user is guided to contact with key functions. For example, the key function may be a broadcast station function for providing a teacher user with a notification, issuing a job, or issuing a card-punching task into a binding class, where a parent user may view the corresponding content within the APP, facilitating communication between the parent and the teacher. However, due to the problem of user habit, the user occupation of the broadcasting station function is relatively low, and the operator hopes to popularize the function, so that a certain intervention measure can be performed on a certain time point of a part of users to guide the users to contact the function.

The processing method of the user behavior provided by the embodiment of the invention can be divided into an engine end and a service server end, wherein the engine end utilizes the historical one-week behavior data of the user to construct a weighted directed graph, and calculates the path from non-key behavior to key behavior in the directed graph in a certain step number and the corresponding weight; the server side acquires a real-time behavior sequence of the user, returns whether the user is a key user needing intervention, and gives the probability of the user touching the key function and predicts the contact path.

FIG. 5 shows a concrete implementation flow of the engine side, as shown in FIG. 5, a batch of key actions needs to be predefined first, that is, functions that are rarely used in the app at present, but the operator wants the user to touch; the engine end obtains historical behavior data of the full quantity of users in the app; after data cleaning is carried out on some predefined noise behaviors, stain behaviors and the like, carrying out behavior sequence cutting on the predefined noise behaviors, stain behaviors and the like, and carrying out behavior sequence cleaning; classifying the sequence into a positive sample and a negative sample according to whether the sequence contains key behaviors, namely, whether the sequence contains the key behaviors is regarded as the positive sample, otherwise, the sequence is classified into the negative sample, and filtering the sequence under certain conditions; simultaneously dividing the training set and the testing set to construct a positive training set and a negative training set and a testing set; constructing a user behavior graph by using a positive sample training set, constructing a feature to calculate the weight of each edge, and then trimming the edge of the network under a certain condition by using a negative sample training set to obtain an optimal user behavior graph; and carrying out conditional path searching on all non-key behaviors to key behaviors in the graph according to the behavior graph, and giving out the weight corresponding to the path.

Fig. 6 shows a specific implementation flow of the engine end construction dataset, firstly constructing positive and negative samples, then randomly selecting samples to construct a training set and a test set, and finally randomly selecting samples of the training set to construct the training set and a verification set.

Fig. 7 shows a specific implementation flow of the server, as shown in fig. 7, the server obtains a real-time behavior sequence of a user, searches whether last behavior in the behavior sequence of the user meeting a condition at a certain time point exists in a given path, if so, judges whether the corresponding weight meets an intervention condition, and if so, returns the behavior path and the weight which are most likely to reach a key function by the user and the corresponding weight high, medium and low levels.

Through the constructed directed graph of the user behaviors, whether the user is the user needing intervention or not can be judged according to the real-time behavior sequence of the user, the future behavior path of the user is predicted, the proper user is guided to use the key function through the proper path, and the user experience and the user viscosity are improved. Meanwhile, the user has more use data on the APP, so that the later user data analysis and mining are facilitated, and better service and virtuous circle are provided for the user.

Example 6

The following are examples of the apparatus of the present application that may be used to perform the method embodiments of the present application. For details not disclosed in the embodiments of the apparatus of the present application, please refer to the embodiments of the method of the present application.

As shown in fig. 8, the processing means of the user behavior may be implemented as all or part of the electronic device by software, hardware or a combination of both. The apparatus includes a first acquisition module 82, a construction module 84, and a retrieval module 86.

The first obtaining module 82 is configured to obtain behavior data of a plurality of objects;

the constructing module 84 is configured to construct a directed graph based on behavior data of a plurality of objects, where the directed graph includes: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges;

the retrieving module 86 is configured to retrieve the directed graph, determine at least one behavior path for transferring the non-target behavior to the target behavior, and a weight value of each behavior path, where the number of different behaviors included in the behavior path is less than or equal to a preset number;

on the basis of the above embodiment, the building module includes: the dividing unit is used for dividing the behavior data of the objects to obtain a plurality of behavior sequences; a first construction unit, configured to construct a training set and a test set based on a plurality of behavior sequences, where the training set includes: a positive sample training set and a negative sample training set, the test set comprising: a positive sample test set and a negative sample test set, wherein the positive sample is a behavior sequence containing target behaviors, and the negative sample is a behavior sequence not containing target behaviors; the second construction unit is used for constructing the first directed graph by utilizing the training set; the testing unit is used for testing the first directed graph based on the testing set to obtain a testing result; and the updating unit is used for updating the first directed graph until the test result meets the preset condition and obtaining the directed graph under the condition that the test result does not meet the preset condition.

On the basis of the above-described embodiment, the second construction unit includes: a building subunit, configured to build a second directed graph using the positive sample training set; a first determining subunit, configured to determine a first behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the positive sample edge-trimming rate; a second determining subunit, configured to determine a second behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the negative-sample edge-trimming rate; and the edge cutting subunit is used for cutting edges of the second directed graph based on the first behavior sequence and the second behavior sequence to obtain a first directed graph.

On the basis of the above-described embodiment, the updating unit includes: the first updating subunit is used for updating the positive sample edge-cutting rate and the negative sample edge-cutting rate based on a preset step length to obtain updated positive sample edge-cutting rate and negative sample edge-cutting rate; a third determining subunit, configured to determine a third behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the updated positive sample edge-trimming rate; a fourth determining subunit, configured to determine a fourth behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the updated negative-sample edge-trimming rate; the edge-cutting subunit is further configured to cut edges of the second directed graph based on the third behavior sequence and the fourth behavior sequence, so as to obtain a third directed graph, where the third directed graph is determined to be the directed graph when a test result corresponding to the second directed graph meets a preset condition.

On the basis of the above embodiment, the constructing subunit includes: the first processing submodule is used for processing the positive sample training set and determining nodes and edges in the second directed graph; the extraction sub-module is used for carrying out feature extraction on the behavior data of the two behaviors connected by the edges to obtain feature data corresponding to the edges; and the second processing sub-module is used for obtaining the weight value of the edge based on the characteristic data corresponding to the edge.

On the basis of the above-described embodiment, the first construction unit includes: a fifth determining subunit for determining a positive sample data set and a negative sample data set based on the plurality of behavior sequences; the first dividing subunit is used for randomly dividing the positive sample data set to obtain a positive sample training set and a positive sample testing set; and the second dividing subunit is used for randomly dividing the negative sample data set to obtain a negative sample training set and a negative sample testing set.

On the basis of the embodiment, the device further comprises: the judging module is used for judging whether the test result meets the preset condition, wherein the judging module comprises: the first determining unit is used for determining a first recognition rate based on a first test result of the positive sample test set, wherein the first recognition rate is used for representing the duty ratio of the positive sample test result in the first test result; the second determining unit is used for determining a second recognition rate based on a second test result of the negative sample test set, wherein the second recognition rate is used for representing the duty ratio of the positive sample test result in the second test result; an obtaining unit, configured to obtain a first variation of the first recognition rate and a second variation of the second recognition rate after updating the positive sample trimming rate and the negative sample trimming rate; a first judging unit for judging whether the first recognition rate is smaller than a preset value, whether the first recognition rate is smaller than the second recognition rate, and whether the first variation is smaller than the second variation; the third determining unit is used for determining that the test result meets the preset condition if the first recognition rate is smaller than the preset value, the first recognition rate is smaller than the second recognition rate, and the first variation is smaller than the second variation; and a fourth determining unit configured to determine that the test result does not satisfy the preset condition if the first recognition rate is greater than or equal to the preset value, the first recognition rate is greater than or equal to the second recognition rate, or the first variation is greater than or equal to the second variation.

On the basis of the above embodiment, the building module further includes: the matching unit is used for matching the behavior data with preset behavior data; the first eliminating unit is used for eliminating the behavior data if the behavior data is successfully matched with the preset behavior data, so as to obtain cleaned behavior data; the segmentation unit is also used for segmenting the washed behavior data to obtain a plurality of behavior sequences.

On the basis of the above embodiment, the building module further includes: the second judging unit is used for judging whether two adjacent behavior sequences are identical or not; the second eliminating unit is used for eliminating any one of the two adjacent behavior sequences if the two adjacent behavior sequences are the same, so as to obtain a cleaned behavior sequence; the first construction unit is also used for constructing a training set and a testing set based on the washed behavior sequence.

On the basis of the embodiment, the device further comprises: the second acquisition module is used for acquiring a target behavior sequence of the target object; the matching module is used for matching the target behavior sequence with at least one behavior path and determining the target behavior path; the third acquisition module is used for acquiring the weight value of the target behavior path; the determining module is used for determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior.

On the basis of the embodiment, the device further comprises: the monitoring module is used for monitoring whether the target object executes target behaviors; and the updating module is used for updating the directed graph based on the target behavior sequence under the condition that the target object does not execute the target behavior.

Optionally, the apparatus further comprises: the matching module is also used for matching the target behavior data of the target behavior sequence with preset behavior data; the rejecting module is used for rejecting the target behavior data to obtain a cleaned target behavior sequence if the target behavior data is successfully matched with the preset behavior data; the matching module is also used for matching the cleaned target behavior sequence with at least one behavior path to determine the target behavior path.

It should be noted that, when the processing apparatus for user behavior provided in the foregoing embodiment performs the processing method for user behavior, only the division of the foregoing functional modules is used as an example, in practical application, the foregoing functional allocation may be performed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to complete all or part of the functions described above. In addition, the processing device for user behavior and the processing method for user behavior provided in the above embodiments belong to the same concept, and the detailed description of the present process is provided in the method embodiment and is not repeated here.

Example 7

As shown in fig. 9, the processing means of the user behavior may be implemented as all or part of the electronic device by software, hardware or a combination of both. The apparatus includes a second acquisition module 92, a matching module 94, a third acquisition module 96, and a determination module 98.

The second obtaining module 92 is configured to obtain a target behavior sequence of the target object;

the matching module 94 is configured to match the target behavior sequence with at least one behavior path, and determine a target behavior path, where the at least one behavior path is a behavior path for transferring a non-target behavior determined by retrieving a directed graph to a target behavior, the number of different behaviors included in the behavior path is less than or equal to a preset number, the directed graph is constructed based on behavior data of a plurality of objects, and the directed graph includes: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges;

The third obtaining module 96 is configured to obtain a weight value of the target behavior path, where the weight value is determined by retrieving the directed graph;

the determining module 98 is configured to determine whether to push target information based on the weight value of the target behavior path, where the target information is used to prompt the target object to execute the target behavior.

Example 8

The embodiment of the present application further provides a computer storage medium, where the computer storage medium may store a plurality of instructions, where the instructions are adapted to be loaded by a processor and execute the method steps of the embodiment shown in fig. 1 to fig. 7, and the specific execution process may refer to the specific description of the embodiment shown in fig. 1 to fig. 7, which is not repeated herein.

The device on which the storage medium resides may be an electronic device.

Example 9

As shown in fig. 10, the electronic device 1000 may include: at least one processor 1001, at least one network interface 1004, a user interface 1003, a memory 1005, at least one communication bus 1002.

Wherein the communication bus 1002 is used to enable connected communication between these components.

The user interface 1003 may include a Display screen (Display) and a Camera (Camera), and the optional user interface 1003 may further include a standard wired interface and a wireless interface.

The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface), among others.

Wherein the processor 1001 may include one or more processing cores. The processor 1001 connects various parts within the overall electronic device 1000 using various interfaces and lines, performs various functions of the electronic device 1000 and processes data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 1005, and invoking data stored in the memory 1005. Alternatively, the processor 1001 may be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), programmable logic array (Programmable Logic Array, PLA). The processor 1001 may integrate one or a combination of several of a central processing unit (Central Processing Unit, CPU), an image processor (Graphics Processing Unit, GPU), and a modem, etc. The CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing the content required to be displayed by the display screen; the modem is used to handle wireless communications. It will be appreciated that the modem may not be integrated into the processor 1001 and may be implemented by a single chip.

The Memory 1005 may include a random access Memory (Random Access Memory, RAM) or a Read-Only Memory (Read-Only Memory). Optionally, the memory 1005 includes a non-transitory computer readable medium (non-transitory computer-readable storage medium). The memory 1005 may be used to store instructions, programs, code, sets of codes, or sets of instructions. The memory 1005 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-described respective method embodiments, etc.; the storage data area may store data or the like referred to in the above respective method embodiments. The memory 1005 may also optionally be at least one storage device located remotely from the processor 1001. As shown in fig. 10, an operating system, a network communication module, a user interface module, and an operating application of the electronic device may be included in the memory 1005, which is one type of computer storage medium.

In the electronic device 1000 shown in fig. 10, the user interface 1003 is mainly used for providing an input interface for a user, and acquiring data input by the user; and the processor 1001 may be configured to invoke an operating application of the electronic device stored in the memory 1005, and specifically perform the following operations:

Acquiring behavior data of a plurality of objects; constructing a directed graph based on behavior data of a plurality of objects, wherein the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges; searching the directed graph, determining at least one behavior path for transferring non-target behaviors to target behaviors, and a weight value of each behavior path, wherein the number of different behaviors contained in the behavior paths is smaller than or equal to a preset number; acquiring a target behavior sequence of a target object; and matching the target behavior sequence with at least one behavior path to determine the target behavior path.

In one embodiment, the operating system of the electronic device is an android system, in which the processor 1001 further performs the following steps:

dividing behavior data of a plurality of objects to obtain a plurality of behavior sequences; constructing a training set and a testing set based on a plurality of behavior sequences, wherein the training set comprises: a positive sample training set and a negative sample training set, the test set comprising: a positive sample test set and a negative sample test set, wherein the positive sample is a behavior sequence containing target behaviors, and the negative sample is a behavior sequence not containing target behaviors; constructing a first directed graph by using the training set; testing the first directed graph based on the test set to obtain a test result; and under the condition that the test result does not meet the preset condition, updating the first directed graph until the test result meets the preset condition, and obtaining the directed graph.

In one embodiment, the processor 1001 further performs the steps of:

determining a first behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the positive sample edge-shearing rate; determining a second behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the negative-sample edge-shearing rate; and cutting edges of the second directed graph based on the first behavior sequence and the second behavior sequence to obtain a first directed graph.

In one embodiment, the processor 1001 further performs the steps of:

updating the positive sample edge shearing rate and the negative sample edge shearing rate based on a preset step length to obtain updated positive sample edge shearing rate and negative sample edge shearing rate; determining a third behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the updated positive sample edge-shearing rate; determining a fourth behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the updated negative-sample edge-shearing rate; and cutting edges of the second directed graph based on the third behavior sequence and the fourth behavior sequence to obtain a third directed graph, wherein the third directed graph is determined to be the directed graph under the condition that a test result corresponding to the second directed graph meets a preset condition.

In one embodiment, the processor 1001 further performs the steps of:

processing the positive sample training set to determine nodes and edges in the second directed graph; extracting characteristics of behavior data of two behaviors connected by edges to obtain characteristic data corresponding to the edges; and obtaining the weight value of the edge based on the characteristic data corresponding to the edge.

In one embodiment, the processor 1001 further performs the steps of:

determining a positive sample data set and a negative sample data set based on the plurality of behavior sequences; randomly dividing a positive sample data set to obtain a positive sample training set and a positive sample testing set; and randomly dividing the negative sample data set to obtain a negative sample training set and a negative sample testing set.

In one embodiment, the processor 1001 further performs the steps of:

determining a first recognition rate based on a first test result of the positive sample test set, wherein the first recognition rate is used for representing the duty ratio of the positive sample test result in the first test result; determining a second recognition rate based on a second test result of the negative sample test set, wherein the second recognition rate is used for representing the duty ratio of the positive sample test result in the second test result; after updating the positive sample trimming rate and the negative sample trimming rate, acquiring a first variation of the first recognition rate and a second variation of the second recognition rate; judging whether the first recognition rate is smaller than a preset value, whether the first recognition rate is smaller than the second recognition rate, and whether the first variation is smaller than the second variation; if the first recognition rate is smaller than the preset value, the first recognition rate is smaller than the second recognition rate, and the first variation is smaller than the second variation, determining that the test result meets the preset condition; if the first recognition rate is greater than or equal to a preset value, the first recognition rate is greater than or equal to the second recognition rate, or the first variation is greater than or equal to the second variation, determining that the test result does not meet the preset condition.

In one embodiment, the processor 1001 further performs the steps of:

before dividing behavior data of a plurality of objects to obtain a plurality of behavior sequences, matching the behavior data with preset behavior data; if the behavior data is successfully matched with the preset behavior data, the behavior data is removed, and the cleaned behavior data is obtained; and dividing the washed behavior data to obtain a plurality of behavior sequences.

In one embodiment, the processor 1001 further performs the steps of:

before a training set and a testing set are built based on a plurality of behavior sequences, judging whether two adjacent behavior sequences are identical or not; if the two adjacent behavior sequences are the same, any one of the two adjacent behavior sequences is eliminated to obtain a cleaned behavior sequence; and constructing a training set and a testing set based on the washed behavior sequence.

In one embodiment, the processor 1001 further performs the steps of:

after the target behavior sequence is matched with at least one behavior path and the target behavior path is determined, the weight value of the target behavior path is obtained; and determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior.

In one embodiment, the processor 1001 further performs the steps of:

after determining to push the target information, monitoring whether the target object executes target behavior; and updating the directed graph based on the target behavior sequence under the condition that the target object does not execute the target behavior.

In one embodiment, the processor 1001 further performs the steps of:

before the target behavior sequence is matched with at least one behavior path and the target behavior path is determined, matching the target behavior data of the target behavior sequence with preset behavior data; if the target behavior data is successfully matched with the preset behavior data, the target behavior data is removed, and a cleaned target behavior sequence is obtained; and matching the cleaned target behavior sequence with at least one behavior path to determine a target behavior path.

The method has the advantages that the weighted directed graph is constructed by utilizing the historical behavior data of the whole user group, and the future behavior path of the user is predicted through the weighted directed graph, so that the technical effects of improving the prediction accuracy, improving the user experience and improving the user viscosity are achieved, and the technical problems of low accuracy and poor user experience of the user behavior prediction in the related technology are solved.

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises an element.

The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.

Claims

1. A method for processing user behavior, comprising:

acquiring behavior data of a plurality of objects, wherein the plurality of objects are users using application programs installed on electronic equipment;

constructing a directed graph based on behavior data of the plurality of objects, wherein the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes and weight values of the edges, wherein the nodes are used for representing different behaviors in the behavior data, and the edges are used for representing transfer relations between two behaviors connected through the edges;

searching the directed graph, determining at least one behavior path for transferring non-target behaviors to target behaviors, and a weight value of each behavior path, wherein the target behaviors are predefined key behaviors, the non-target behaviors are other behaviors except the target behaviors, and the number of different behaviors contained in the behavior paths is smaller than or equal to a preset number;

Based on behavior data of the plurality of objects, constructing a directed graph includes: dividing behavior data of the objects to obtain a plurality of behavior sequences;

constructing a training set and a testing set based on the plurality of behavior sequences, wherein the training set comprises: a positive sample training set and a negative sample training set, the test set comprising: a positive sample test set and a negative sample test set, wherein the positive sample is a behavior sequence containing the target behavior, and the negative sample is a behavior sequence not containing the target behavior;

constructing a first directed graph by using the training set;

testing the first directed graph based on the test set to obtain a test result;

under the condition that the test result does not meet the preset condition, updating the first directed graph until the test result meets the preset condition, and obtaining the directed graph;

wherein, judging whether the test result meets the preset condition comprises: determining a first recognition rate based on a first test result of the positive sample test set, wherein the first recognition rate is used for representing the duty ratio of the positive sample test result in the first test result;

Determining a second recognition rate based on a second test result of the negative sample test set, wherein the second recognition rate is used for representing the duty ratio of a positive sample test result in the second test result;

after updating the positive sample edge-cutting rate and the negative sample edge-cutting rate, acquiring a first variation of the first recognition rate and a second variation of the second recognition rate;

judging whether the first recognition rate is smaller than a preset value, whether the first recognition rate is smaller than the second recognition rate, and whether the first variation is smaller than the second variation;

if the first recognition rate is smaller than the preset value, the first recognition rate is smaller than the second recognition rate, and the first variation is smaller than the second variation, determining that the test result meets the preset condition;

and if the first recognition rate is greater than or equal to the preset value, the first recognition rate is greater than or equal to the second recognition rate, or the first variation is greater than or equal to the second variation, determining that the test result does not meet the preset condition.

2. The method of claim 1, wherein constructing an initial directed graph using the training set comprises:

Constructing a second directed graph by using the positive sample training set;

determining a first behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the positive sample edge-shearing rate;

determining a second behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the negative-sample edge-shearing rate;

and cutting edges of the second directed graph based on the first behavior sequence and the second behavior sequence to obtain the first directed graph.

3. The method of claim 2, wherein updating the first directed graph comprises:

updating the positive sample edge shearing rate and the negative sample edge shearing rate based on a preset step length to obtain updated positive sample edge shearing rate and negative sample edge shearing rate;

determining a third behavior sequence in the positive sample training set based on the weight value of each edge in the second directed graph and the updated positive sample edge-trimming rate;

determining a fourth behavior sequence in the negative-sample training set based on the weight value of each edge in the second directed graph and the updated negative-sample edge-trimming rate;

and cutting edges of the second directed graph based on the third behavior sequence and the fourth behavior sequence to obtain a third directed graph, wherein the third directed graph is determined to be the directed graph under the condition that a test result corresponding to the second directed graph meets the preset condition.

4. The method of claim 2, wherein constructing a second directed graph using the positive sample training set comprises:

processing the positive sample training set, and determining nodes and edges in the second directed graph;

performing feature extraction on behavior data of two behaviors connected by the edge to obtain feature data corresponding to the edge;

and obtaining the weight value of the edge based on the characteristic data corresponding to the edge.

5. The method of claim 1, wherein constructing a training set and a test set based on the plurality of behavior sequences comprises:

determining a positive sample data set and a negative sample data set based on the plurality of behavior sequences;

randomly dividing the positive sample data set to obtain the positive sample training set and the positive sample testing set;

and randomly dividing the negative sample data set to obtain the negative sample training set and the negative sample testing set.

6. The method of claim 1, wherein prior to partitioning the behavioral data of the plurality of objects to obtain a plurality of behavioral sequences, the method further comprises:

matching the behavior data with preset behavior data;

If the behavior data are successfully matched with the preset behavior data, eliminating the behavior data to obtain cleaned behavior data;

and dividing the washed behavior data to obtain the behavior sequences.

7. The method of claim 1, wherein prior to constructing the training set and the test set based on the plurality of behavior sequences, the method further comprises:

judging whether two adjacent behavior sequences are identical;

if the two adjacent behavior sequences are the same, any one of the two adjacent behavior sequences is removed to obtain a washed behavior sequence;

and constructing the training set and the testing set based on the cleaned behavior sequence.

8. The method according to any one of claims 1 to 7, wherein after retrieving the directed graph, determining at least one behavior path for non-target behavior to transfer to target behavior, and a weight value for each behavior path, the method further comprises:

acquiring a target behavior sequence of a target object;

matching the target behavior sequence with the at least one behavior path to determine a target behavior path;

Acquiring a weight value of the target behavior path;

and determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior.

9. The method of claim 8, wherein after determining to push the target information, the method further comprises:

monitoring whether the target object executes the target behavior;

and updating the directed graph based on the target behavior sequence under the condition that the target object does not execute the target behavior.

10. The method of claim 8, wherein prior to matching the target behavior sequence with the at least one behavior path, determining a target behavior path, the method further comprises:

matching the target behavior data of the target behavior sequence with preset behavior data;

if the target behavior data is successfully matched with the preset behavior data, eliminating the target behavior data to obtain a cleaned target behavior sequence;

and matching the cleaned target behavior sequence with the at least one behavior path to determine the target behavior path.

11. A method for processing user behavior, comprising:

acquiring a target behavior sequence of a target object;

matching the target behavior sequence with at least one behavior path to determine a target behavior path, wherein the at least one behavior path is a behavior path for transferring non-target behaviors determined by searching a directed graph to target behaviors, the number of different behaviors contained in the behavior path is smaller than or equal to a preset number, the directed graph is constructed based on behavior data of a plurality of objects, and the directed graph comprises: the system comprises a plurality of nodes, edges connected between any two nodes, and weight values of the edges, wherein the nodes are used for representing different behaviors in behavior data, the edges are used for representing a transfer relation between two behaviors connected through the edges, the target behavior is a predefined key behavior, the non-target behavior is other behaviors except the target behavior, and the plurality of objects are users using application programs installed on electronic equipment;

acquiring a weight value of the target behavior path, wherein the weight value is determined by searching the directed graph;

Determining whether to push target information based on the weight value of the target behavior path, wherein the target information is used for prompting the target object to execute the target behavior;

constructing a first directed graph by using the training set;

testing the first directed graph based on the test set to obtain a test result;

12. A computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method steps of any of claims 1 to 11.

13. An electronic device, comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the method steps of any of claims 1 to 11.