Anchor point pushing method and device based on user behavior analysis, equipment and medium

Abstract

The invention provides an anchor point pushing method and device based on user behavior analysis, equipment and a medium, and the method comprises the steps: obtaining a behavior record of a user, generating a fingerprint, determining a target anchor point of a first user, and pushing the information of a product historically purchased by a second user to the first user based on the target anchor point. The method has the beneficial effects that the similarity between users is fully utilized, the effect of personalized recommendation is realized, the commodity exposure rate and conversion rate can be effectively improved, the sales growth is promoted, the manual intervention and operation cost of enterprises can be remarkably reduced through automatic and intelligent user behavior analysis and product information pushing process, and in addition, the user experience is improved. User behavior changes can be tracked and analyzed in real time, and recommendation strategies are adjusted in time to cope with market changes and dynamic evolution of user requirements.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to an anchor point push method, device, equipment and medium based on user behavior analysis. Background Art

[0002] In the prior art, the method of recommending products to users is generally based on the user's search content. However, for information technology, users are not clear about the specific parameters of the information, which makes it impossible to choose information technology products, so recommendations are needed.

[0003] With the rapid development of the Internet, user behavior data is recorded and stored, including but not limited to user browsing history, search keywords, click behavior, purchasing habits, geographic location information, social interaction, etc. The existing technology does not have a way to recommend products to users based on user behavior data. Therefore, there is an urgent need for an anchor push method based on user behavior analysis. Summary of the invention

[0004] The main purpose of the present invention is to provide an anchor push method, device, equipment and medium based on user behavior analysis, aiming to solve the problem that the prior art has no way to recommend products to users based on user behavior data.

[0005] The present invention provides an anchor point push method based on user behavior analysis, comprising:

[0006] Obtaining a first behavior record of a first user and a second behavior record of a second user;

[0007] The first user generates a first behavior fingerprint based on the first behavior record, and the second user generates a second behavior fingerprint based on the second behavior record;

[0008] The first user broadcasts based on the first behavior fingerprint, and the second user broadcasts based on the second behavior record;

[0009] obtaining a similarity between the first behavior fingerprint and the second behavior fingerprint, and when the similarity between the first behavior fingerprint and the second behavior fingerprint is greater than a preset similarity, the first user receiving a second behavior record of the second user;

[0010] Analyzing a target anchor point of the first user based on the first behavior record;

[0011] The product information historically purchased by the second user is pushed to the first user based on the target anchor point.

[0012] Furthermore, the step of pushing the product information of the second user's historical purchases to the first user based on the target anchor point includes:

[0013] According to the anchor point, two linear functions are set using a preset linear classifier. Among them, b t =b t-1 +m t , and b 1 =m 1 , m t represents the correlation constant corresponding to the anchor point, b t represents the bias, t is a positive integer, w represents the weight vector, f t (x) represents a linear function, x represents an anchor point, and W is a preset parameter;

[0014] Obtain the product information of the second user's historical purchases, calculate the Euclidean distance between each linear function and each product information, extract the maximum Euclidean distance and the minimum Euclidean distance of each linear function, and use the difference between the two as the behavior distance of the corresponding linear function;

[0015] According to the formula Calculate the transformation parameter A of the behavioral distance between two linear functions i ; where t n represents the nth behavior distance, T(t n ) indicates that based on t n The preset calculation function of

[0016] Determine the transformation parameter A i Is it within the preset range?

[0017] If it is within a preset range, multiple target product information is selected from the product information historically purchased by the second user based on two linear functions to form push information to be pushed to the first user.

[0018] Furthermore, the determination of the transformation parameter A i After the step of whether it is within the preset range, it also includes:

[0019] If it is not within the preset range, adjusting the value of the weight vector in the linear function until the transformation parameter is within the preset range, thereby obtaining two target linear functions;

[0020] Based on two target linear functions, multiple target product information is selected from the product information historically purchased by the second user to form push information recommended to the first user.

[0021] Furthermore, before the step of the first user receiving the second behavior record of the second user, the method further includes:

[0022] Collecting evaluation content of each purchased product information in the second behavior record of the second user; wherein the evaluation content corresponds to the product information one by one;

[0023] Using sentiment analysis tools to analyze the evaluation content to obtain words representing sentiment tendencies and their sentiment polarity values;

[0024] Counting the probability of occurrence of words with positive emotions and the probability of occurrence of words with negative emotions in the evaluation content; wherein, words with a sentiment polarity value greater than a set value are words with positive emotions, and words with a sentiment polarity value less than or equal to the set value are words with negative emotions;

[0025] Obtaining the emotion score of the second user for each product information according to the ratio of the probability of occurrence of positive emotion words to the probability of occurrence of negative emotion words;

[0026] Determine whether the number of product information whose sentiment scores reach a preset score is greater than a preset number;

[0027] If it is greater than a preset number, it is determined that the second behavior record of the second user can be shared.

[0028] Furthermore, before the step of the first user receiving the second behavior record of the second user, the method further includes:

[0029] Determining whether there is a third user that satisfies that a third behavior fingerprint of the third user matches the first behavior fingerprint of the first user;

[0030] If yes, calculating the sentiment score of the third behavior record of the third user;

[0031] Determining whether the sentiment score of the third behavior record of the third user is greater than the sentiment score of the second behavior record of the second user;

[0032] If so, the third behavior record of the third user is selected and sent to the first user.

[0033] Furthermore, the step of analyzing the target anchor point of the first user based on the first behavior record includes:

[0034] Preprocessing the first behavior record to obtain a standard first behavior record; wherein the preprocessing includes data cleaning, data conversion and data normalization of the first behavior record;

[0035] The standard first behavior record is processed by a preset machine learning algorithm to identify the user's behavior pattern;

[0036] Extracting potential anchor points of the user according to the behavior pattern;

[0037] Analyzing the correlation between the potential anchor point and other behaviors of the user;

[0038] A target anchor point among the potential anchor points is determined according to the magnitude of the association.

[0039] Further, after the step of pushing the product information of the second user's historical purchases to the first user based on the target anchor point, the method further includes:

[0040] Determining whether the first user has generated new behavior information;

[0041] If new behavior information is generated, the first behavior fingerprint is updated based on the new behavior information.

[0042] The present invention also provides an anchor point push device based on user behavior analysis, comprising:

[0043] An acquisition module, used to acquire a first behavior record of a first user and a second behavior record of a second user;

[0044] A generating module, configured for the first user to generate a first behavior fingerprint based on the first behavior record, and for the second user to generate a second behavior fingerprint based on the second behavior record;

[0045] A broadcast module, configured for the first user to broadcast based on the first behavior fingerprint, and for the second user to broadcast based on the second behavior record;

[0046] a receiving module, configured to obtain a similarity between the first behavior fingerprint and the second behavior fingerprint, and when the similarity between the first behavior fingerprint and the second behavior fingerprint is greater than a preset similarity, the first user receives the second behavior record of the second user;

[0047] An analysis module, configured to analyze a target anchor point of the first user based on the first behavior record;

[0048] A push module is used to push the product information of the second user's history purchase to the first user based on the target anchor point.

[0049] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0050] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0051] The beneficial effects of the present invention are as follows: by obtaining the user's behavior record and generating a fingerprint, the target anchor point of the first user is determined, and based on the target anchor point, the product information of the second user's historical purchase is pushed to the first user. The similarity between users is fully utilized to achieve the effect of personalized recommendation. For e-commerce platforms or retailers, the anchor point push method based on user behavior analysis can effectively improve the product exposure rate and conversion rate, thereby driving sales growth. At the same time, this method can also help to tap potential users and market opportunities and create more commercial value for enterprises. Through the automated and intelligent user behavior analysis and product information push process, enterprises can significantly reduce manual intervention and operating costs. In addition, the method can also track and analyze user behavior changes in real time, and adjust the recommendation strategy in time to cope with market changes and the dynamic evolution of user needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flowchart of an anchor point push method based on user behavior analysis according to an embodiment of the present invention;

[0053] Figure 2 It is a schematic block diagram of the structure of an anchor point push device based on user behavior analysis according to an embodiment of the present invention;

[0054] Figure 3 A schematic block diagram of the structure of a computer device according to an embodiment of the present application.

[0055] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0056] 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.

[0057] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly. The connection can be a direct connection or an indirect connection.

[0058] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone.

[0059] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0060] Reference Figure 1 The present invention proposes an anchor point push method based on user behavior analysis, comprising:

[0061] S1: Obtain a first behavior record of a first user and a second behavior record of a second user;

[0062] S2: The first user generates a first behavior fingerprint based on the first behavior record, and the second user generates a second behavior fingerprint based on the second behavior record;

[0063] S3: The first user broadcasts based on the first behavior fingerprint, and the second user broadcasts based on the second behavior record;

[0064] S4: obtaining a similarity between the first behavior fingerprint and the second behavior fingerprint, and when the similarity between the first behavior fingerprint and the second behavior fingerprint is greater than a preset similarity, the first user receives the second behavior record of the second user;

[0065] S5: Analyze the target anchor point of the first user based on the first behavior record;

[0066] S6: Pushing product information historically purchased by the second user to the first user based on the target anchor point.

[0067] As described in the above step S1, a first behavior record of the first user and a second behavior record of the second user are obtained, wherein the first behavior record of the first user and the second behavior record of the second user can be obtained in a preset platform database, and the first behavior record and the second behavior record may include various online activity data of users such as browsing, clicking, purchasing, and commenting.

[0068] As described in step S2 above, the first user generates a first behavior fingerprint based on the first behavior record, and the second user generates a second behavior fingerprint based on the second behavior record. The first user and the second user generate unique behavior fingerprints based on their respective behavior records. A behavior fingerprint is a highly generalized representation that can reflect the user's behavioral characteristics and preferences. The process of generating a behavior fingerprint can specifically be to convert the corresponding behavior record into a corresponding feature, and then vectorize the feature to generate a corresponding feature vector, thereby obtaining the corresponding behavior fingerprint.

[0069] As described in step S3 above, the first user broadcasts based on the first behavior fingerprint, and the second user broadcasts based on the second behavior record. Broadcasting is to publish the behavior fingerprint to the system so that other users have the opportunity to match and compare with it. The broadcasting method can be broadcasting in the corresponding APP or through a wireless device, which can facilitate each user to quickly find other matching users.

[0070] As described in step S4 above, the similarity between the first behavior fingerprint and the second behavior fingerprint is obtained, and when the similarity between the first behavior fingerprint and the second behavior fingerprint is greater than the preset similarity, the first user receives the second behavior record of the second user. The system compares the behavior fingerprints of the first user and the second user, and when the similarity between the two is greater than the preset similarity, it is considered that the two users have similar behavioral characteristics and preferences. At this time, the first user will receive the second behavior record of the second user to further understand the interests and purchasing habits of the second user. Among them, the method of calculating the similarity is not limited, for example, it can be calculated by calculating cosine similarity, Euclidean distance, etc. The first user can receive the second behavior record of the second user. Of course, in order to protect user privacy, sensitive information in the second behavior record can be deleted before sending it to the first user.

[0071] As described in step S5 above, the target anchor point of the first user is analyzed based on the first behavior record. Based on the first behavior record of the first user, the system analyzes the target anchor point of the first user. The target anchor point can be understood as the product category, brand, functional point, etc. that the first user is currently most concerned about or most interested in. Among them, the acquisition of the target anchor point can be obtained based on the analysis of the first behavior record of the first user. Specifically, it can be identified through semantic recognition technology, or keywords or behaviors in the first behavior record can be extracted and then identified. It should be noted that the target anchor point is the information of the ideal product corresponding to the user.

[0072] As described in step S6 above, the product information of the second user's historical purchases is pushed to the first user based on the target anchor point. Specifically, the recommendation method can be to recommend the product information purchased by the second user to the first user. In some embodiments, the second user may include multiple users, and the product information purchased by the second user may also be multiple. Therefore, the correlation between each product information and the target anchor point can be calculated to determine the pushed product. This push method makes full use of the similarity between users and achieves the effect of personalized recommendation. For e-commerce platforms or retailers, the anchor point push method based on user behavior analysis can effectively improve the exposure rate and conversion rate of goods, thereby driving sales growth. At the same time, this method also helps to tap potential users and market opportunities and create more commercial value for enterprises. Through automated and intelligent user behavior analysis and product information push processes, enterprises can significantly reduce manual intervention and operating costs. In addition, the method can also track and analyze user behavior changes in real time, and adjust the recommendation strategy in time to cope with market changes and the dynamic evolution of user needs.

[0073] In one embodiment, the step S6 of pushing the product information historically purchased by the second user to the first user based on the target anchor point includes:

[0074] S601: Using a preset linear classifier to set two linear functions according to the anchor point Among them, b t =b t-1 +m t , and b 1 =m 1 , m t represents the correlation constant corresponding to the anchor point, b t represents the bias, t is a positive integer, w represents the weight vector, f t (x) represents a linear function, x represents an anchor point, and W is a preset parameter;

[0075] S602: Obtain the product information of the second user's historical purchases, calculate the Euclidean distance between each linear function and each product information, extract the maximum Euclidean distance and the minimum Euclidean distance of each linear function, and use the difference between the two as the behavior distance of the corresponding linear function;

[0076] S603: According to the formula Calculate the transformation parameter A of the behavioral distance between two linear functions i ; where t n represents the nth behavior distance, T(t n ) indicates that based on t n The preset calculation function of

[0077] S604: Determine the transformation parameter A iIs it within the preset range?

[0078] S605: If it is within a preset range, multiple target product information is selected from the product information historically purchased by the second user based on two linear functions to form push information to be pushed to the first user.

[0079] As described in the above step S601, two linear functions are set according to the anchor point using a preset linear classifier, wherein the preset linear classifier can be any linear classifier of SVM (linear kernel), linear regression, Bayesian classification, single-layer perceptron, and LR, which can be regarded as a hyperplane to separate the samples. Setting two linear functions is equivalent to setting two hyperplanes for classification to screen out the target product information.

[0080] As described in the above steps S602-S604, the Euclidean distance between each linear function and each product information is calculated, and the maximum Euclidean distance and the minimum Euclidean distance of each linear function are extracted, and the difference between the two is used as the behavioral distance of the corresponding linear function. The larger the behavioral distance, the farther it is from the ideal product information of the first user, and the larger the behavioral distance, the closer it is to the ideal product information of the first user. Then, by calculating the transformation parameters, when the transformation parameters are closer to 1, the better the classification effect of the two linear functions, otherwise, the worse the classification effect. When the classification effect is better, the corresponding screening accuracy will be higher. Therefore, it can be determined whether the design of the two linear functions is reasonable by judging whether the transformation parameters are within the preset range. If the setting is reasonable, multiple target product information is selected from the product information historically purchased by the second user based on the two linear functions to form push information to the first user.

[0081] In one embodiment, the determining of the transformation parameter A i After step S603 of determining whether the data is within a preset range, the method further includes:

[0082] S6041: If it is not within the preset range, adjusting the value of the weight vector in the linear function until the transformation parameter is within the preset range, thereby obtaining two target linear functions;

[0083] S6042: Select multiple target product information from the product information historically purchased by the second user based on two target linear functions to form push information recommended to the first user.

[0084] As described in the above steps S6041-S6042, the adjustment of the target linear function is achieved. When the transformation parameter A iWhen it is not within the preset range, it means that the classification of the two linear functions is unreasonable. Therefore, the value of the weight vector in the linear function can be adjusted. The adjustment method is not limited. In some embodiments, the corresponding adjustment can be made according to the result of the transformation parameter. For example, if the transformation parameter is far away from the preset range, a large adjustment can be made. If the transformation parameter is far away from the preset range, a small adjustment can be made. The adjustment range is reasonably planned until the transformation parameter is within the preset range, thereby obtaining two target linear functions. Based on the two target linear functions, multiple target product information is selected from the product information historically purchased by the second user to form push information recommended to the first user. The selection method is the same as above and will not be repeated here.

[0085] In one embodiment, before step S4 in which the first user receives the second behavior record of the second user, the method further includes:

[0086] S301: Collecting evaluation content of each purchased product information in the second behavior record of the second user; wherein the evaluation content corresponds to the product information one by one;

[0087] S302: Analyze the evaluation content using a sentiment analysis tool to obtain words representing sentiment tendencies and their sentiment polarity values;

[0088] S303: Counting the probability of occurrence of words with positive emotions and the probability of occurrence of words with negative emotions in the evaluation content; wherein, words with a sentiment polarity value greater than a set value are words with positive emotions, and words with a sentiment polarity value less than or equal to the set value are words with negative emotions;

[0089] S304: Obtaining the emotion score of the second user for each product information according to the ratio of the probability of occurrence of positive emotion words to the probability of occurrence of negative emotion words;

[0090] S305: Determine whether the number of product information whose sentiment scores reach a preset score is greater than a preset number;

[0091] S306: If the number is greater than a preset number, it is determined that the second behavior record of the second user can be shared.

[0092] As described in the above steps S301-S306, further screening of the second user is achieved. Specifically, the sentiment analysis tool SentiWordNet (SentiWordNet can divide the evaluation content into positivity and negativity according to the sentiment score) can be used to analyze the comment content to obtain the words representing the sentiment tendency and their sentiment polarity values. The sentiment polarity value can be set by itself. For example, the words with a sentiment polarity value greater than 0.3 are set as positive sentiment words, and the words with a sentiment polarity value less than or equal to 0.3 are set as negative sentiment words. The probability of positive sentiment words and the probability of negative sentiment words in the comment content are further counted. The sentiment score of the second user for each product information is obtained according to the ratio of the probability of positive sentiment words to the probability of negative sentiment words. If the number of virtual carriers with a preset sentiment score is greater than the preset number, it is determined that the second reading record of the second user can be shared. It should be noted that in order to facilitate the first user to screen products with good usability after purchase, this embodiment further screens the second user, so that the product information recommended to the first user can better meet the needs of the first user.

[0093] In one embodiment, before step S4 of the first user receiving the second behavior record of the second user, the method further includes:

[0094] S311: Determine whether there is a third user satisfying that the third behavior fingerprint of the third user matches the first behavior fingerprint of the first user;

[0095] S312: If yes, calculating the sentiment score of the third behavior record of the third user;

[0096] S313: Determine whether the emotion score of the third behavior record of the third user is greater than the emotion score of the second behavior record of the second user;

[0097] S314: If yes, select the third behavior record of the third user and send it to the first user.

[0098] As described in the above steps S311-S314, during the broadcasting process, there may be multiple users who match the behavioral fingerprint of the first user. When the emotional score of the third behavior record of the third user is calculated to be higher, it means that the third user may be more satisfied with the product. Therefore, the third behavior record of the third user is selected and sent to the first user. Subsequently, the first user is recommended through the product information of the third user. The method of selecting product information is the same as mentioned above and will not be repeated here.

[0099] In one embodiment, the step S5 of analyzing the target anchor point of the first user based on the first behavior record includes:

[0100] S501: preprocessing the first behavior record to obtain a standard first behavior record; wherein the preprocessing includes data cleaning, data conversion and data normalization of the first behavior record;

[0101] S502: performing recognition processing on the standard first behavior record by using a preset machine learning algorithm to identify a user's behavior pattern;

[0102] S503: extracting potential anchor points of the user according to the behavior pattern;

[0103] S504: Analyze the correlation between the potential anchor point and other behaviors of the user;

[0104] S505: Determine a target anchor point among the potential anchor points according to the magnitude of the correlation.

[0105] As described in step S501 above, the first behavior record is preprocessed to obtain a standard first behavior record. Specifically, noise and invalid data in the behavior record, such as duplicate records, erroneous data, or irrelevant information, can be removed by data cleaning. This helps to ensure the accuracy and efficiency of subsequent analysis. Non-numeric data is converted into numerical data (such as converting text classification labels into numeric codes), or data is converted into a format suitable for processing by machine learning algorithms. Normalizing the data can ensure that all features are on the same scale, thereby avoiding certain features from dominating the algorithm, and scaling the data to a specific range (0 to 1).

[0106] As described in step S502 above, the recognition process is performed by a preset machine learning algorithm. The pre-processed behavior records are analyzed using a pre-selected machine learning algorithm (such as a clustering, classification or regression algorithm). The purpose of the algorithm is to identify the user's behavior patterns, which may include the user's preferences, habits or behavior trends.

[0107] As described in step S503 above, potential anchor points of the user are extracted, wherein potential anchor points refer to key factors or events that have a significant impact on user behavior. These anchor points can be identified through preset algorithm analysis, such as semantic recognition, or a custom recognition method.

[0108] As described in step S504 above, the correlation between the potential anchor point and other behaviors of the user is analyzed. The causal relationship or correlation between the potential anchor point and other behaviors of the user is explored. The calculation method of the correlation is not limited, for example, it can be the Pearson correlation coefficient, the Spearman rank correlation coefficient, the Kendall rank correlation coefficient, etc. Through statistical analysis and data mining technology, it can be revealed how these anchor points affect other behaviors of the user.

[0109] As described in step S505 above, the target anchor point is determined according to the magnitude of the correlation. The target anchor point is screened out from the potential anchor points according to the strength and importance of the correlation. These target anchor points are the factors that have the greatest impact on user behavior and are the most representative. They are crucial for understanding user behavior and formulating corresponding strategies.

[0110] In one embodiment, after the step S6 of pushing the product information of the second user's historical purchases to the first user based on the target anchor point, the method further includes:

[0111] S701: Determine whether the first user has generated new behavior information;

[0112] S702: If new behavior information is generated, update the first behavior fingerprint based on the new behavior information.

[0113] As described in the above steps S701-S702, the first behavior fingerprint is updated. Since the user will generate new behavior information during the purchase process, the behavior fingerprint needs to be updated according to the new behavior information. The updated behavior fingerprint will be used for future personalized recommendations and marketing strategies to ensure that the system can continue to accurately reflect the user's latest preferences and needs. This enables the system to keep real-time tracking of user behavior and dynamically adjust the recommendation strategy based on the user's latest behavior. This helps to improve user satisfaction and purchase conversion rate because the system can more accurately understand the user's needs and interests.

[0114] The beneficial effects of the present invention are as follows: by obtaining the user's behavior record and generating a fingerprint, the target anchor point of the first user is determined, and based on the target anchor point, the product information of the second user's historical purchase is pushed to the first user. The similarity between users is fully utilized to achieve the effect of personalized recommendation. For e-commerce platforms or retailers, the anchor point push method based on user behavior analysis can effectively improve the product exposure rate and conversion rate, thereby driving sales growth. At the same time, this method can also help to tap potential users and market opportunities and create more commercial value for enterprises. Through the automated and intelligent user behavior analysis and product information push process, enterprises can significantly reduce manual intervention and operating costs. In addition, the method can also track and analyze user behavior changes in real time, and adjust the recommendation strategy in time to cope with market changes and the dynamic evolution of user needs.

[0115] Reference Figure 2 The present invention also provides an anchor point push device based on user behavior analysis, comprising:

[0116] An acquisition module 10 is used to acquire a first behavior record of a first user and a second behavior record of a second user;

[0117] A generating module 20, configured for the first user to generate a first behavior fingerprint based on the first behavior record, and for the second user to generate a second behavior fingerprint based on the second behavior record;

[0118] A broadcast module 30, configured for the first user to broadcast based on the first behavior fingerprint, and for the second user to broadcast based on the second behavior record;

[0119] The receiving module 40 is used to obtain the similarity between the first behavior fingerprint and the second behavior fingerprint, and when the similarity between the first behavior fingerprint and the second behavior fingerprint is greater than a preset similarity, the first user receives the second behavior record of the second user;

[0120] An analysis module 50, configured to analyze a target anchor point of the first user based on the first behavior record;

[0121] The push module 60 is used to push the product information of the second user's historical purchases to the first user based on the target anchor point.

[0122] In one embodiment, the push module 60 includes:

[0123] The linear function setting submodule is used to set two linear functions according to the anchor point using a preset linear classifier. Among them, b t =b t-1 +m t , and b 1 =m 1 , m t represents the correlation constant corresponding to the anchor point, b t represents the bias, t is a positive integer, w represents the weight vector, f t (x) represents a linear function, x represents an anchor point, and W is a preset parameter;

[0124] A product information acquisition submodule, used to acquire the product information of the second user's historical purchases, calculate the Euclidean distance between each linear function and each product information, extract the maximum Euclidean distance and the minimum Euclidean distance of each linear function, and use the difference between the two as the behavior distance of the corresponding linear function;

[0125] The transformation parameter calculation submodule is used to calculate the Calculate the transformation parameter A of the behavioral distance between two linear functions i ; where t n represents the nth behavior distance, T(t n ) indicates that based on t n The preset calculation function of

[0126] The transformation parameter judgment submodule is used to judge the transformation parameter Ai Is it within the preset range?

[0127] The target product information acquisition submodule is used to select multiple target product information from the product information purchased historically by the second user based on two linear functions if it is within a preset range, so as to form push information to be pushed to the first user.

[0128] In one embodiment, the push module 60 further includes:

[0129] A weight vector adjustment module, used for adjusting the value of the weight vector in the linear function if it is not within a preset range, until the transformation parameter is within a preset range, thereby obtaining two target linear functions;

[0130] The target product information selection submodule is used to select multiple target product information from the product information historically purchased by the second user based on two target linear functions to form push information recommended to the first user.

[0131] In one embodiment, the anchor point push device based on user behavior analysis further includes:

[0132] An evaluation content collection module, used to collect evaluation content of each purchased product information in the second behavior record of the second user; wherein the evaluation content corresponds to the product information one by one;

[0133] An evaluation content analysis module is used to analyze the evaluation content using a sentiment analysis tool to obtain words representing sentiment tendencies and their sentiment polarity values;

[0134] A probability statistics module, used to count the probability of occurrence of words with positive emotions and the probability of occurrence of words with negative emotions in the evaluation content; wherein, words with a sentiment polarity value greater than a set value are words with positive emotions, and words with a sentiment polarity value less than or equal to the set value are words with negative emotions;

[0135] An emotion score calculation module, used to obtain the emotion score of the second user for each product information according to the ratio of the probability of occurrence of positive emotion words to the probability of occurrence of negative emotion words;

[0136] The emotion molecule judgment module is used to judge whether the number of product information with emotion scores reaching a preset score is greater than a preset number;

[0137] The determination module is used to determine that the second behavior record of the second user can be shared if the number is greater than a preset number.

[0138] Furthermore, the anchor point push device based on user behavior analysis also includes:

[0139] A behavior fingerprint judgment module, used to judge whether there is a third user satisfying that a third behavior fingerprint of the third user matches the first behavior fingerprint of the first user;

[0140] a sentiment score calculation module, configured to calculate the sentiment score of the third behavior record of the third user;

[0141] A sentiment score determination module, used to determine whether the sentiment score of the third behavior record of the third user is greater than the sentiment score of the second behavior record of the second user;

[0142] The third behavior record sending module is used to select the third behavior record of the third user and send it to the first user.

[0143] Furthermore, the analysis module 50 includes:

[0144] A preprocessing submodule, configured to preprocess the first behavior record to obtain a standard first behavior record; wherein the preprocessing includes data cleaning, data conversion and data normalization of the first behavior record;

[0145] An identification processing submodule, used to identify the standard first behavior record through a preset machine learning algorithm to identify the user's behavior pattern;

[0146] An extraction submodule, used for extracting potential anchor points of the user according to the behavior pattern;

[0147] A correlation line analysis submodule, used to analyze the correlation between the potential anchor point and other behaviors of the user;

[0148] The target anchor point determination submodule is used to determine the target anchor point among the potential anchor points according to the magnitude of the association.

[0149] Furthermore, the anchor point push device based on user behavior analysis also includes:

[0150] A new behavior information determination module, used to determine whether the first user has generated new behavior information;

[0151] An updating module is used to update the first behavior fingerprint based on new behavior information if new behavior information is generated.

[0152] Reference Figure 3 In an embodiment of the present application, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store various behavior records, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the anchor point push method based on user behavior analysis described in any of the above embodiments can be implemented.

[0153] Those skilled in the art will understand that Figure 3 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0154] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the anchor point push method based on user behavior analysis described in any of the above embodiments can be implemented.

[0155] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0156] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, device, article or method including the element.

[0157] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0158] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0159] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the claims of the present invention.

Claims

1. An anchor point push method based on user behavior analysis, characterized in that: include: Obtaining a first behavior record of a first user and a second behavior record of a second user; The first user generates a first behavior fingerprint based on the first behavior record, and the second user generates a second behavior fingerprint based on the second behavior record; The first user broadcasts based on the first behavior fingerprint, and the second user broadcasts based on the second behavior record; obtaining a similarity between the first behavior fingerprint and the second behavior fingerprint, and when the similarity between the first behavior fingerprint and the second behavior fingerprint is greater than a preset similarity, the first user receiving a second behavior record of the second user; Analyzing a target anchor point of the first user based on the first behavior record; The product information historically purchased by the second user is pushed to the first user based on the target anchor point.

2. The anchor point push method based on user behavior analysis according to claim 1, characterized in that: The step of pushing the product information of the second user's historical purchases to the first user based on the target anchor point includes: According to the anchor point, two linear functions are set using a preset linear classifier. Among them, b t =b t-1 +m t , and b1=m1,m t represents the correlation constant corresponding to the anchor point, b t represents the bias, t is a positive integer, w represents the weight vector, f t (x) represents a linear function, x represents an anchor point, and W is a preset parameter; Obtain the product information of the second user's historical purchases, calculate the Euclidean distance between each linear function and each product information, extract the maximum Euclidean distance and the minimum Euclidean distance of each linear function, and use the difference between the two as the behavior distance of the corresponding linear function; According to the formula Calculate the transformation parameter A of the behavioral distance between two linear functions i ; where t n represents the nth behavior distance, T(t n ) indicates that based on t n The preset calculation function of Determine the transformation parameter A i Is it within the preset range? If it is within a preset range, multiple target product information is selected from the product information historically purchased by the second user based on two linear functions to form push information to be pushed to the first user.

3. The anchor point push method based on user behavior analysis according to claim 2, characterized in that: The determination of the transformation parameter A i After the step of whether it is within the preset range, it also includes: If it is not within the preset range, adjusting the value of the weight vector in the linear function until the transformation parameter is within the preset range, thereby obtaining two target linear functions; Based on two target linear functions, multiple target product information is selected from the product information historically purchased by the second user to form push information recommended to the first user.

4. The anchor point push method based on user behavior analysis according to claim 1, characterized in that: Before the step of the first user receiving the second behavior record of the second user, the method further includes: Collecting evaluation content of each purchased product information in the second behavior record of the second user; wherein the evaluation content corresponds to the product information one by one; Using sentiment analysis tools to analyze the evaluation content to obtain words representing sentiment tendencies and their sentiment polarity values; Counting the probability of occurrence of words with positive emotions and the probability of occurrence of words with negative emotions in the evaluation content; wherein, words with a sentiment polarity value greater than a set value are words with positive emotions, and words with a sentiment polarity value less than or equal to the set value are words with negative emotions; Obtaining the emotion score of the second user for each product information according to the ratio of the probability of occurrence of positive emotion words to the probability of occurrence of negative emotion words; Determine whether the number of product information whose sentiment scores reach a preset score is greater than a preset number; If it is greater than a preset number, it is determined that the second behavior record of the second user can be shared.

5. The anchor point push method based on user behavior analysis according to claim 4, characterized in that: Before the step of the first user receiving the second behavior record of the second user, the method further includes: Determining whether there is a third user that satisfies that a third behavior fingerprint of the third user matches the first behavior fingerprint of the first user; If yes, calculating the sentiment score of the third behavior record of the third user; Determining whether the sentiment score of the third behavior record of the third user is greater than the sentiment score of the second behavior record of the second user; If so, the third behavior record of the third user is selected and sent to the first user.

6. The anchor point push method based on user behavior analysis according to claim 1, characterized in that: The step of analyzing the target anchor point of the first user based on the first behavior record includes: Preprocessing the first behavior record to obtain a standard first behavior record; wherein the preprocessing includes data cleaning, data conversion and data normalization of the first behavior record; The standard first behavior record is processed by a preset machine learning algorithm to identify the user's behavior pattern; Extracting potential anchor points of the user according to the behavior pattern; Analyzing the correlation between the potential anchor point and other behaviors of the user; A target anchor point among the potential anchor points is determined according to the magnitude of the association.

7. The anchor point push method based on user behavior analysis according to claim 1, characterized in that: After the step of pushing the product information of the second user's historical purchases to the first user based on the target anchor point, the method further includes: Determining whether the first user has generated new behavior information; If new behavior information is generated, the first behavior fingerprint is updated based on the new behavior information.

8. An anchor point push device based on user behavior analysis, characterized in that: include: An acquisition module, used to acquire a first behavior record of a first user and a second behavior record of a second user; A generating module, configured for the first user to generate a first behavior fingerprint based on the first behavior record, and for the second user to generate a second behavior fingerprint based on the second behavior record; A broadcast module, configured for the first user to broadcast based on the first behavior fingerprint, and for the second user to broadcast based on the second behavior record; a receiving module, configured to obtain a similarity between the first behavior fingerprint and the second behavior fingerprint, and when the similarity between the first behavior fingerprint and the second behavior fingerprint is greater than a preset similarity, the first user receives the second behavior record of the second user; An analysis module, configured to analyze a target anchor point of the first user based on the first behavior record; A push module is used to push the product information of the second user's history purchase to the first user based on the target anchor point.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images