A behavior activity analysis method and system

By collecting and analyzing entertainment activity data, determining user interest points and combining explicit and implicit association relationships, using a fully connected network for interest points recommendation, solving the problem of failing to effectively utilize entertainment activity data in the existing technology, and achieving the latest update High accuracy and personalized recommendations of interest enrich users' entertainment life.

CN114692010BActive Publication Date: 2024-12-03SUINING WANDERING TECH CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202210196977.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-01
Publication Date
2024-12-03
Estimated Expiration
2042-03-01

AI Technical Summary

Technical Problem

The existing technology has failed to effectively and deeply utilize the data of entertainment activities, resulting in insufficient evaluation and enrichment of users' cultural and entertainment life.

Method used

By collecting historical entertainment activity data, extracting interest indicator data, determining the user's enthusiasm factor level for entertainment activities, analyzing the characteristics of interest points, combining explicit and implicit relationships, using a fully connected network for interest points recommendation.

Benefits of technology

It improves the accuracy and personalization of point-of-interest recommendations, helps users match similar activities, expand social circles, and enrich cultural and entertainment life.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114692010B_ABST
    Figure CN114692010B_ABST
Patent Text Reader

Abstract

The present invention discloses a behavior activity analysis method and system, which collects historical entertainment activity data information; extracts interest index data from the entertainment activity data information, converts the interest index data into standard index values, and then fills them into a pre-set index text to form an interest index set; determines the level of the current user's active factors closely related to various entertainment activities according to the maximum value of the ratio of each interest index corresponding to the entertainment activity data information; determines the corresponding interest points based on the level of the active factors and the interest index set; classifies the interest points, analyzes the relevant characteristics of each type of interest points, and mines the coupling relationship between user-interest points; inputs the user-interest point coupling relationship into a fully connected network to recommend interest points; the above scheme realizes the evaluation of behavior activities in entertainment, and has a high accuracy in recommending interest points.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a method and a system, and in particular to a behavior activity analysis method and a system. Background Art

[0002] Since 1978, my country's entertainment industry has gone through more than 40 years of development, and has gone through the start-up stage, primary development stage, rapid development stage and industrial explosion stage. Now that people's living standards have improved, they will participate in various entertainment activities to enrich their lives, but the market has not made in-depth use of such data after users have participated in entertainment activities. Summary of the invention

[0003] In order to solve the above problems, the present invention provides a behavior activity analysis method and system, which realizes the evaluation of behavior activities in entertainment, has high accuracy in recommending points of interest, effectively helps users match like-minded activities, expands users' social circles, and enriches users' entertainment life.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A behavior activity analysis method, the method comprising:

[0006] Collect data and information on historical cultural and entertainment activities;

[0007] Extracting interest index data from the entertainment activity data information, converting the interest index data into standard index values ​​and then filling them into a pre-set index text to form an interest index set;

[0008] Determine the level of the current user's enthusiasm factors closely related to various entertainment activities according to the maximum value of the ratio of various interest indicators corresponding to the entertainment activity data information;

[0009] Determining corresponding points of interest based on the levels of the motivation factors and the set of interest indicators;

[0010] Classify points of interest, analyze the relevant features of various points of interest, and explore the coupling relationship between users and points of interest;

[0011] The user-POI coupling relationship is input into a fully connected network to recommend POIs.

[0012] Preferably, the entertainment activity data information includes the activity content, activity location, form of expression, social information, user comment information on hot spots, and ways to participate in entertainment activities;

[0013] The various interest indicators corresponding to the entertainment activity data information include: entertainment activity participation rate, talent award rate and grade level certification rate.

[0014] Preferably, the step of determining the level of the positive factors closely related to each entertainment activity according to the maximum value of the ratio of each interest index corresponding to the entertainment activity data information comprises:

[0015]

[0016] Among them, w represents the maximum value of the ratio of each interest index corresponding to the entertainment activity data information, x 1 ,x 2 ,x 3 ,x 4 There are four levels that represent the positive factors closely related to various cultural and recreational activities, namely, very fond of, moderately fond of, fond of, and generally fond of.

[0017] Preferably, the step of extracting interest indicator data from the entertainment activity data information, converting the interest indicator data into standard indicator values ​​and then filling the values ​​into a pre-set indicator text to form an interest indicator set includes:

[0018] According to the activity content, activity location, form of expression, and ways of participating in the cultural and entertainment activities, corresponding interest indicator data is obtained, wherein the interest indicator data is marked with an interest indicator category, one indicator category corresponds to one interest indicator, and each interest indicator contains at least one sub-level indicator;

[0019] Convert the interest indicator data of different indicator categories into standard indicator values, determine the sub-level indicators corresponding to each standard indicator value, write the standard indicator value into the indicator text corresponding to each sub-level indicator, and connect the indicator texts of each sub-level indicator to obtain a single-category indicator text;

[0020] Connect all individual indicator texts to form a set of interest indicators.

[0021] Furthermore, converting the interest indicator data into a standard indicator value includes:

[0022] determining an order of magnitude based on the interest indicator data;

[0023] The interest indicator data is processed by digitization according to the order of magnitude, and a unit word corresponding to the order of magnitude is added to obtain a standard indicator value.

[0024] Preferably, the classification of interest points and analysis of relevant features of each type of interest points include: using a K-means algorithm to cluster the interest points so that interest points with similar content are clustered into one category, and each type of interest point has a corresponding class label for identifying an explicit association between the interest point and the user's attributes.

[0025] Preferably, mining the coupling relationship between users and points of interest includes: learning the explicit association relationship between users and points of interest in attributes through a pre-established convolutional neural network model;

[0026] Based on the user's check-in data at the POI, learn the implicit association between the user and the POI;

[0027] The explicit and implicit associations between users and POIs are integrated to characterize the coupling relationship between users and POIs.

[0028] Furthermore, the clustering of interest points using the K-means algorithm includes:

[0029] Randomly select k interest points from the interest point set as the initial cluster centers; calculate the distance between each interest point and each cluster center, and assign each interest point to the cluster center closest to it. The cluster centers and the interest points assigned to them represent a cluster. The cluster center of each cluster is recalculated based on the existing interest points in the cluster until the termination condition of K-means clustering is met.

[0030] Furthermore, the Euclidean distance ρ from the remaining interest points to the cluster center is calculated by the following formula, and the closest interest points are placed into the category to form a new class;

[0031]

[0032] Among them, v i.λ =(lat i ,lon i ) and v j.λ =(lat j ,lon j ) are interest point datasets V = {v 1 ,v 2 ,…,v n}Two points of interest in .

[0033] A behavior activity analysis system, the system comprising:

[0034] The collection module is used to collect data information of historical entertainment activities;

[0035] A processing module, used for extracting interest index data from the entertainment activity data information, converting the interest index data into standard index values ​​and then filling them into a pre-set index text to form an interest index set;

[0036] The first determination module is used to determine the level of the current user's enthusiasm factors closely related to various entertainment activities according to the maximum value of the ratio of various interest indicators corresponding to the entertainment activity data information;

[0037] A second determination module, configured to determine corresponding points of interest based on the level of the positivity factor and the set of interest indicators;

[0038] Mining module, used to classify points of interest, analyze the relevant features of various points of interest, and mine the coupling relationship between users and points of interest;

[0039] The recommendation module is used to input the user-POI coupling relationship into a fully connected network to recommend POIs.

[0040] Technical effects and advantages of the present invention:

[0041] The embodiment of the present invention provides a behavior activity analysis method and system, which collects historical entertainment activity data information, extracts interest index data from the entertainment activity data information and the level of user's enthusiasm for participating in various entertainment activities, thereby determining interest points and effectively analyzing the interest points. This fills the gap in the existing methods and approaches for evaluating entertainment activity data.

[0042] The coupling relationship between users and POIs is comprehensively considered, and the explicit and implicit associations between users and POIs are integrated. Through a convolutional neural network model, the explicit association between user attributes and POI attributes is captured. Compared with the existing technology, the POI recommendation proposed in this paper has higher accuracy and personalization, and achieves better recommendation effect. It effectively helps users match activities with like-minded people, expands their social circles, and enriches their entertainment life. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0044] Figure 1 The present invention is a flow chart of a behavior activity analysis method. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] Embodiment 1: In this specific embodiment, a behavior activity analysis method is provided, such as Figure 1As shown, the method includes:

[0047] S1 collects data and information on historical entertainment activities;

[0048] S2 extracts interest index data from the entertainment activity data information, converts the interest index data into standard index values, and then fills them into a pre-set index text to form an interest index set;

[0049] S3 determines the level of the current user's enthusiasm factors closely related to various entertainment activities according to the maximum value of the ratio of various interest indicators corresponding to the entertainment activity data information;

[0050] S4, determining corresponding points of interest based on the level of the positivity factor and the set of interest indicators;

[0051] S5 classifies points of interest, analyzes the relevant features of various points of interest, and mines the coupling relationship between users and points of interest;

[0052] S6 inputs the user-POI coupling relationship into a fully connected network to recommend POIs.

[0053] In step S1, the historical entertainment activity data information collected includes activity content, activity location, form of expression, social information, user comment information on hot spots, ways to participate in entertainment activities, etc.; the collection method includes submitting performance viewing credentials and location information of the performance being attended.

[0054] In step S2, various interest indicators corresponding to the extracted entertainment activity data information include: entertainment activity participation rate, talent award rate and grade level certification rate, etc.

[0055] Step S2 extracts the interest index data in the entertainment activity data information, converts the interest index data into standard index values ​​and fills them into a pre-set index text to form an interest index set, including:

[0056] According to the activity content, activity location, form of expression, and ways of participating in the cultural and entertainment activities, corresponding interest indicator data is obtained, wherein the interest indicator data is marked with an interest indicator category, one indicator category corresponds to one interest indicator, and each interest indicator contains at least one sub-level indicator;

[0057] Convert the interest indicator data of different indicator categories into standard indicator values, determine the sub-level indicators corresponding to each standard indicator value, write the standard indicator value into the indicator text corresponding to each sub-level indicator, and connect the indicator texts of each sub-level indicator to obtain a single-category indicator text;

[0058] Connect all individual indicator texts to form a set of interest indicators.

[0059] The converting of the interest indicator data into a standard indicator value includes:

[0060] determining an order of magnitude based on the interest indicator data;

[0061] The interest indicator data is processed by digitization according to the order of magnitude, and a unit word corresponding to the order of magnitude is added to obtain a standard indicator value.

[0062] In step S3, according to the maximum value of the ratio of each interest index corresponding to the entertainment activity data information, the level of the active factors closely related to each entertainment activity is determined, including:

[0063]

[0064] Among them, w represents the maximum value of the ratio of each interest index corresponding to the entertainment activity data information, x 1 ,x 2 ,x 3 ,x 4 There are four levels that represent the positive factors closely related to various cultural and recreational activities, namely, very fond of, moderately fond of, fond of, and generally fond of.

[0065] In step S5, the points of interest are classified and the relevant features of each type of points of interest are analyzed, including: using the K-means algorithm to cluster the points of interest so that points of interest with similar content are clustered into one category, and each type of point of interest has a corresponding class label for identifying an explicit association between the point of interest and the user's attributes.

[0066] Among them, the K-means algorithm is used to cluster the points of interest including:

[0067] Randomly select k interest points from the interest point set as the initial cluster centers; calculate the distance between each interest point and each cluster center, and assign each interest point to the cluster center closest to it. The cluster centers and the interest points assigned to them represent a cluster. The cluster center of each cluster is recalculated based on the existing interest points in the cluster until the termination condition of K-means clustering is met.

[0068] The processing process of interest point clustering based on K-means algorithm is as follows:

[0069] 1) Randomly select k interest points from the interest point set as the initial cluster centers.

[0070] 2) Calculate the Euclidean distance ρ from the remaining interest points to the cluster center by the following formula, put the closest interest points into the category, and form a new class;

[0071]

[0072] Among them, v i.λ =(lat i ,lon i ) and v j.λ =(lat j ,lon j ) are interest point datasets V = {v 1 ,v 2 ,…,v n}Two points of interest in .

[0073] 3) Take the mean of the longitude and latitude of all points of interest in the current cluster as the new center point, and update the point of interest closest to the cluster center;

[0074] 4) Until the objective function converges or the cluster center remains unchanged, otherwise it will transfer to step 2).

[0075] 5) Output the interest point clustering results.

[0076] In step S6, mining the coupling relationship between users and interest points includes: learning the explicit association relationship between users and interest points in attributes through a pre-established convolutional neural network model;

[0077] Based on the user's check-in data at the POI, learn the implicit association between the user and the POI;

[0078] The explicit and implicit associations between users and POIs are integrated to characterize the coupling relationship between users and POIs.

[0079] Among them, the convolutional neural network model consists of convolutional layers and pooling layers. The convolutional layers and pooling layers can extract deeper features. The execution method is:

[0080] a p =pooling(g(w*x c +b))

[0081] Where: * is the convolution operator; w is the filter; b is the bias of w; g is the nonlinear activation function; pooling is the pooling function (for example, maximum pooling or average pooling).

[0082] On this basis, the user-POI association matrix is ​​flattened into a vector as a global explicit association vector, and it is connected with the local explicit association vector to comprehensively represent the explicit association relationship between users and POIs.

[0083] Embodiment 2:

[0084] Based on the same technical concept, a behavior activity analysis system is also provided in a specific embodiment of the present invention, including:

[0085] The collection module is used to collect data information of historical entertainment activities;

[0086] A processing module, used for extracting interest index data from the entertainment activity data information, converting the interest index data into standard index values ​​and then filling them into a pre-set index text to form an interest index set;

[0087] The first determination module is used to determine the level of the current user's enthusiasm factors closely related to various entertainment activities according to the maximum value of the ratio of various interest indicators corresponding to the entertainment activity data information;

[0088] A second determination module, configured to determine corresponding points of interest based on the level of the positivity factor and the set of interest indicators;

[0089] Mining module, used to classify points of interest, analyze the relevant features of various points of interest, and mine the coupling relationship between users and points of interest;

[0090] The recommendation module is used to input the user-POI coupling relationship into a fully connected network to recommend POIs.

[0091] Finally, it should be noted that although the above embodiments have been described with reference to the preferred embodiments, various modifications may be made thereto and components thereof may be replaced with equivalents without departing from the scope of the present invention. In particular, the various technical features mentioned in the various embodiments may be combined in any manner as long as there is no structural conflict. The present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A behavior activity analysis method, characterized in that: The method comprises: Collect data and information on historical cultural and entertainment activities; Extracting interest index data from the entertainment activity data information, converting the interest index data into standard index values ​​and then filling them into a pre-set index text to form an interest index set; Determine the level of the current user's enthusiasm factors closely related to various entertainment activities according to the maximum value of the ratio of various interest indicators corresponding to the entertainment activity data information; Determining corresponding points of interest based on the levels of the motivation factors and the set of interest indicators; Classify points of interest, analyze the relevant features of various points of interest, and explore the coupling relationship between users and points of interest; Input the user-POI coupling relationship into a fully connected network to recommend POIs; The entertainment activity data information includes the activity content, activity location, form of expression, social information, user comments on hot spots, and ways to participate in entertainment activities; The various interest indicators corresponding to the entertainment activity data information include: entertainment activity participation rate, talent award rate and grade level certification rate; The step of extracting interest indicator data from the entertainment activity data information, converting the interest indicator data into standard indicator values ​​and then filling them into a pre-set indicator text to form an interest indicator set includes: According to the activity content, activity location, form of expression, and ways of participating in the cultural and entertainment activities, corresponding interest indicator data is obtained, wherein the interest indicator data is marked with an interest indicator category, one indicator category corresponds to one interest indicator, and each interest indicator contains at least one sub-level indicator; Convert the interest indicator data of different indicator categories into standard indicator values, determine the sub-level indicators corresponding to each standard indicator value, write the standard indicator value into the indicator text corresponding to each sub-level indicator, and connect the indicator texts of each sub-level indicator to obtain a single-category indicator text; Connect all individual indicator texts to form a set of interest indicators; The mining of the coupling relationship between users and points of interest includes: learning the explicit association relationship between users and points of interest in attributes through a pre-established convolutional neural network model; Based on the user's check-in data at the POI, learn the implicit association between the user and the POI; Integrate the explicit and implicit associations between users and POIs to represent the coupling relationship between users and POIs; The user-POI association matrix is ​​flattened into a vector as a global explicit association vector, which is then connected with the local explicit association vector to comprehensively represent the explicit association relationship between the user and POI.

2. The method according to claim 1, characterized in that Determining the level of the positive factors closely related to each entertainment activity according to the maximum value of the ratio of each interest index corresponding to the entertainment activity data information includes: Among them, w represents the maximum value of the ratio of various interest indicators corresponding to the entertainment activity data information, x1, x2, x3, x4 represent the four levels of positive factors closely related to various entertainment activities, namely, very fond of, moderately fond of, fond of, and general fond of.

3. The method according to claim 2, characterized in that The converting the interest indicator data into a standard indicator value comprises: determining an order of magnitude based on the interest indicator data; The interest indicator data is processed by digitization according to the order of magnitude, and a unit word corresponding to the order of magnitude is added to obtain a standard indicator value.

4. The method according to claim 1, characterized in that: The classification of interest points and analysis of relevant features of each type of interest points include: using the K-means algorithm to cluster the interest points so that interest points with similar content are clustered into one category, and each type of interest point has a corresponding class label for identifying an explicit association between the interest point and the user's attributes.

5. The method according to claim 4, characterized in that The K-means algorithm is used to cluster the points of interest, including: Randomly select k interest points from the interest point set as the initial cluster centers; calculate the distance between each interest point and each cluster center, and assign each interest point to the cluster center closest to it. The cluster centers and the interest points assigned to them represent a cluster. The cluster center of each cluster is recalculated based on the existing interest points in the cluster until the termination condition of K-means clustering is met.

6. The method according to claim 5, characterized in that The Euclidean distance ρ from the remaining interest points to the cluster center is calculated by the following formula, and the closest interest points are placed into the category to form a new class; Among them, v i.λ =(lat i ,lon i ) and v j.λ =(lat j ,lon j ) are interest point datasets V = {v1, v2, …, v n }Two points of interest in .

7. A behavior activity analysis system, characterized in that: The system comprises: The collection module is used to collect data information of historical entertainment activities; A processing module, used for extracting interest index data from the entertainment activity data information, converting the interest index data into standard index values ​​and then filling them into a pre-set index text to form an interest index set; The first determination module is used to determine the level of the current user's enthusiasm factors closely related to various entertainment activities according to the maximum value of the ratio of various interest indicators corresponding to the entertainment activity data information; A second determination module, configured to determine corresponding points of interest based on the level of the positivity factor and the set of interest indicators; Mining module, used to classify points of interest, analyze the relevant features of various points of interest, and mine the coupling relationship between users and points of interest; The recommendation module is used to input the user-POI coupling relationship into a fully connected network to recommend POIs; The entertainment activity data information includes the activity content, activity location, form of expression, social information, user comments on hot spots, and ways to participate in entertainment activities; The various interest indicators corresponding to the entertainment activity data information include: entertainment activity participation rate, talent award rate and grade level certification rate; The step of extracting interest indicator data from the entertainment activity data information, converting the interest indicator data into standard indicator values ​​and then filling them into a pre-set indicator text to form an interest indicator set includes: According to the activity content, activity location, form of expression, and ways of participating in the cultural and entertainment activities, corresponding interest indicator data is obtained, wherein the interest indicator data is marked with an interest indicator category, one indicator category corresponds to one interest indicator, and each interest indicator contains at least one sub-level indicator; Convert the interest indicator data of different indicator categories into standard indicator values, determine the sub-level indicators corresponding to each standard indicator value, write the standard indicator value into the indicator text corresponding to each sub-level indicator, and connect the indicator texts of each sub-level indicator to obtain a single-category indicator text; Connect all individual indicator texts to form a set of interest indicators; The mining of the coupling relationship between users and points of interest includes: learning the explicit association relationship between users and points of interest in attributes through a pre-established convolutional neural network model; Based on the user's check-in data at the POI, learn the implicit association between the user and the POI; Integrate the explicit and implicit associations between users and POIs to represent the coupling relationship between users and POIs; The user-POI association matrix is ​​flattened into a vector as a global explicit association vector, which is then connected with the local explicit association vector to comprehensively represent the explicit association relationship between the user and POI.

Citation Information

Patent Citations

  • Interest point recommendation method fusing contextual features of user continuous sign-in behavior sequence

    CN112801749A