Content recommendation method, device and equipment and computer readable storage medium

Abstract

The invention provides a content recommendation method, device and equipment and a computer readable storage medium. The method comprises the steps of obtaining user behavior data; analyzing the user behavior data to obtain key influence factors; similarity calculation is carried out based on the user behavior data and the key influence factors, a similarity calculation result is obtained, and the similarity calculation result represents the similarity between the users; and based on the similarity calculation result and the key influence factors, generating a content recommendation result by using a recommendation algorithm. According to the method, changes of personal demands can be responded in time by analyzing the user behavior data, so that recommended contents become more flexible and dynamic, real demands and preferences of the user can be more accurately understood by combining the user behavior data and key influence factors and comprehensively considering background and deep factors of the user behaviors, and the user experience is improved. More accurate and more relevant contents are matched for the user, and the satisfaction degree of the user is improved.

Description

Technical Field

[0001] The present application relates to the technical field of content recommendation, and in particular to a content recommendation method, apparatus, device, and computer-readable storage medium. Background Art

[0002] With the development of Internet technology, a large number of applications can provide users with personalized recommendation services and content that suits their behavioral preferences, etc., in order to increase user stickiness and activity.

[0003] Currently, a common approach to content recommendation is to analyze user behavior on a terminal or through an app account to determine user preferences and provide recommended content. However, user behavior on a terminal or through an app account is relatively simple, resulting in relatively concentrated scenarios for the behavioral data obtained, making it difficult to accurately determine user preferences. Furthermore, user preferences can change over time and in the external environment, making traditional static analysis methods difficult to adapt to such changes. Consequently, content recommendation using this approach can easily lead to poor content recommendation accuracy. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a content recommendation method, apparatus, device, and computer-readable storage medium that can more accurately understand the user's real needs and preferences and match more accurate content to the user. The specific technical solution is as follows:

[0005] In a first aspect, the present application provides a content recommendation method, the method comprising:

[0006] Obtain user behavior data;

[0007] Analyze the user behavior data to obtain key influencing factors;

[0008] Performing similarity calculation based on the user behavior data and the key influencing factors to obtain a similarity calculation result, wherein the similarity calculation result represents the similarity between users;

[0009] Based on the similarity calculation result and the key influencing factors, a recommendation algorithm is used to generate a content recommendation result.

[0010] In one possible implementation, analyzing the user behavior data to obtain key influencing factors includes:

[0011] The user behavior data is analyzed using a data lineage analysis method to obtain the key influencing factors.

[0012] In one possible implementation, the method of analyzing the user behavior data using a data lineage analysis method to obtain the key influencing factors includes:

[0013] Analyzing the correlation between the user behavior data based on the user behavior data;

[0014] The user behavior pattern is identified based on the correlation between the user behavior data to obtain the key influencing factors.

[0015] In a possible implementation, performing similarity calculation based on the user behavior data and the key influencing factors to obtain a similarity calculation result includes:

[0016] Determining a weight coefficient for each service based on the key influencing factors, wherein the service is determined based on the user behavior data;

[0017] Calculate the weighted average value of each user's service and the weighted value of each service based on the user behavior data and the weight coefficients of each service;

[0018] A similarity calculation is performed based on the service weighted average value of each user and the service weighted value of each service to obtain a similarity calculation result.

[0019] In a possible implementation, after generating a content recommendation result using a recommendation algorithm based on the similarity calculation result and the key influencing factors, the method further includes:

[0020] Obtaining feedback data for the content recommendation result;

[0021] The recommendation algorithm is optimized according to the feedback data.

[0022] In a second aspect, the present application further provides a content recommendation device, comprising:

[0023] Acquisition module, used to obtain user behavior data;

[0024] An analysis module, configured to analyze the user behavior data to obtain key influencing factors;

[0025] a calculation module, configured to perform similarity calculation based on the user behavior data and the key influencing factors to obtain a similarity calculation result, wherein the similarity calculation result represents the similarity between users;

[0026] A generation module is used to generate content recommendation results using a recommendation algorithm based on the similarity calculation result and the key influencing factors.

[0027] In a possible implementation, the analysis module is specifically configured to:

[0028] The user behavior data is analyzed using a data lineage analysis method to obtain the key influencing factors.

[0029] In a possible implementation, the calculation module includes:

[0030] a weight determination unit, configured to determine a weight coefficient of each service based on the key influencing factors, wherein the service is determined based on the user behavior data;

[0031] A weighted calculation unit, configured to calculate a weighted average value of services for each user and a weighted value of each service based on the user behavior data and the weight coefficients of the services;

[0032] The calculation unit is configured to perform similarity calculation based on the weighted average value of the services of each user and the weighted value of the services of each service to obtain a similarity calculation result.

[0033] In a third aspect, the present application further provides a computer device, comprising: a memory and a processor;

[0034] Wherein, the memory is used to store computer programs;

[0035] The processor is configured to execute the computer program in the memory to implement the first aspect or the method described in any one of the first aspects.

[0036] In a fourth aspect, the present application further provides a computer-readable storage medium storing instructions, which, when executed on a computer, enables the computer to execute the method described in the first aspect or any one of the first aspects.

[0037] In this application, user behavior data is obtained; the user behavior data is analyzed to obtain key influencing factors; similarity calculation is performed based on the user behavior data and the key influencing factors to obtain a similarity calculation result, which represents the similarity between users; based on the similarity calculation result and the key influencing factors, a recommendation algorithm is used to generate content recommendation results. By analyzing user behavior data, this application can respond to changes in individual needs in a timely manner, making recommended content more flexible and dynamic. By combining user behavior data and key influencing factors, it comprehensively considers the background and underlying factors of user behavior, and can more accurately understand the user's true needs and preferences. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0039] Figure 1 A flow chart of a content recommendation method provided by an embodiment of the present application is shown;

[0040] Figure 2 A flowchart of another content recommendation method provided by an embodiment of the present application is shown;

[0041] Figure 3 A schematic structural diagram of a content recommendation device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0042] To make the purpose, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application are within the scope of protection of this application.

[0043] First, some terms that may appear in the embodiments of this application are explained.

[0044] 1) Data Lineage Analysis: This refers to tracing the path of data from its source to its destination, including the origin of the data, how the data is modified by various systems and processes, and how the data is ultimately stored and used. In the embodiments of this application, data lineage analysis is used to identify and understand the generation, transformation, and flow of user behavior data to reveal the motivations and influencing factors behind user behavior.

[0045] 2) The Pearson Correlation Coefficient is a statistical indicator that measures the degree of linear correlation between two sets of data. The value of the Pearson Correlation Coefficient ranges from -1 to 1, where 1 indicates a perfect positive correlation, -1 indicates a perfect negative correlation, and 0 indicates no correlation. In the embodiments of this application, the Pearson Correlation Coefficient is used to represent the similarity between user usage behaviors of different services in an application.

[0046] 3) User Behavior Analysis involves collecting and evaluating user activity within a specific application or system to identify usage patterns, preferences, and behavioral trends. In this embodiment, the user behavior analysis method analyzes the transaction records, query history, and service usage frequency of users using the application to provide personalized marketing services.

[0047] 4) Personalized Recommendation: This is the process of using user behavior data to predict the products or services most likely to attract a specific user. In this embodiment, based on user behavior analysis and similarity calculation, customized products and services are provided to users for online marketing.

[0048] 5) Data sparsity refers to a large amount of data missing or not recorded. In the embodiment of the present application, data sparsity means that users may have less interaction or rating data on certain application services.

[0049] 6) Variable Diversity refers to the different types of data sources and variables that need to be considered in data analysis. In the embodiments of this application, it involves analyzing various factors that affect user behavior in the application, such as economic conditions, life events, market trends, etc.

[0050] Online marketing scenarios primarily address the precise matching of "time + scenario + customer group + behavior / attribute + service / product." This means matching customers with specific attributes or behavior requirements with appropriate services or products at the appropriate time and in the appropriate scenario, in order to achieve the marketing goals of "acquiring, activating, and retaining customers." Precision marketing and precise matching are premised on understanding users. A common approach to understanding users is to collect relevant data such as user behavior and historical transactions, analyze this data, and then group and label users. Finally, matching appropriate services or products to users with corresponding customer groups and tags achieves the purpose of precision marketing. Therefore, it's not difficult to see that accurately analyzing user behavior data is a prerequisite for precision marketing.

[0051] At present, user behavior data is collected in most scenarios in the application. However, due to the "28 principle", that is, 80% of users mainly use 20% of the scenarios in the application, although the average daily collected user behavior data is large in volume, the scenarios corresponding to the data are relatively concentrated, and the data elements are not unified, resulting in problems of data sparsity and variable diversity.

[0052] Current content recommendation technologies mainly rely on traditional statistical methods (such as analyzing user preferences through usage frequency and user ratings), basic machine learning models (such as decision trees, random forests and other algorithms to analyze user behavior), and simple correlation analysis (such as using the Pearson correlation coefficient to analyze the similarity between users). Although these methods are effective to a certain extent, they all face challenges when processing complex data. Challenges include: users may only be active on a limited number of services, resulting in insufficient user behavior data for certain services, which affects the accuracy of the analysis; user behavior is affected by multiple complex factors, and a single data analysis method cannot accurately capture and explain the complex relationships between these factors; user preferences may change over time and the external environment, and traditional static analysis methods cannot adapt to such changes.

[0053] In order to better mine the common characteristics of users from user behavior data and provide a reference for user grouping and stratification when conducting marketing in different application scenarios, the embodiment of the present application provides a content recommendation method, which deeply analyzes the source, flow and conversion process of user behavior data through data lineage analysis, more comprehensively understands user behavior, identifies the potential motivations and factors of user behavior, uses the Pearson correlation coefficient to measure the similarity between users, and deeply explores user behavior patterns. By combining data lineage analysis and the Pearson correlation coefficient, not only the direct performance of user behavior is considered, but also the deep-seated factors and motivations behind the behavior are deeply explored, providing a more dynamic and comprehensive user behavior analysis framework, laying the foundation for subsequent matching strategies and precision marketing, and being able to provide more accurate and personalized products or services to improve user experience and business efficiency.

[0054] See Figure 1 , shows a flow chart of a content recommendation method provided by an embodiment of the present application, which includes at least the following steps:

[0055] S11, obtain user behavior data.

[0056] When content recommendations are needed, initial user behavior data is collected from the app. This initial user behavior data corresponds to the various services within the app. In an app, a service typically refers to a component or module that provides specific functionality or meets specific user needs. These services can be based on different technical architectures and protocols, all designed to enhance user experience and meet business needs.

[0057] Initial user behavior data includes but is not limited to user transaction records (such as account transfers, payment records, loan applications, etc.), search history (such as key search terms), query and browsing data (such as users querying product information, browsing services), user feedback data, login frequency, service usage patterns (such as click streams, page visit time, etc.).

[0058] Currently, the initial user behavior data collection is typically completed by integrating a relevant collection software development kit (SDK) into the application client and calling the relevant collection SDK in various application scenarios. It should be noted that the data collected in this embodiment of the application will be in accordance with the requirements of relevant laws and regulations, and will strictly comply with relevant laws and regulations to ensure that the user's personal information is not abused, and the user's consent will be obtained before collection.

[0059] After collecting the initial user behavior data, the initial user behavior data is preprocessed. Data preprocessing can include data cleaning, data formatting, and data summarization. Data cleaning can include removing invalid or erroneous data, deleting or merging duplicate data, scaling data to a specific range, etc. Data formatting can include converting data into a unified format or unit, and data summarization can include grouping different user behavior data of the same user under the user identity document (ID) of the user. The initial user behavior data that has undergone data preprocessing can be stored in temporary storage or a data lake.

[0060] The initial user behavior data that has undergone data preprocessing is integrated into a unified data framework, and the association relationship between the data is established to facilitate more comprehensive analysis.

[0061] Specifically, Kafka can be used to transmit pre-processed initial user behavior data from temporary storage or a data lake to a central processing system or data warehouse, such as Flink, in real time. Using Flink for real-time data processing and analysis, multi-source, heterogeneous initial user behavior data can be integrated into a unified data framework. This integrated data can then be used for in-depth analysis and personalized services.

[0062] Kafka is widely used as a high-performance message queue system. Within this system, Kafka acts as a message middleware, allowing producers (message senders) to asynchronously send messages to the queue, and consumers (message receivers) to pull messages from the queue for processing on demand. Flink is a distributed real-time big data processing framework that enables real-time analysis of user behavior data.

[0063] Data processing refers to converting pre-processed, raw user behavior data into a format suitable for specific analysis or applications. This can include operations such as data aggregation, transformation, and enrichment. In practice, this may involve merging data from different sources, for example, linking user query data with transaction data; creating derived columns from the data, for example, calculating a user's transaction frequency based on transaction data; and enriching the data by applying business logic, for example, adding initial user clustering tags based on user behavior data. The processed data is richer and more structured, ready for direct use in data analysis or business decision-making.

[0064] After data preprocessing and integration, the initial user behavior data is generated. This data is stored in a database system to ensure its security and accessibility. A database system is a software system specifically designed to store, manage, and retrieve large amounts of data. Examples of database systems include CliHouse and ES.

[0065] When storing user behavior data, you can create indexes for key data fields. Key data fields such as user ID, activity ID, and sequence number can significantly improve query performance, making data retrieval faster and thus improving data analysis efficiency.

[0066] When pre-processing the initial user behavior data, the embodiments of the present application may also encrypt and desensitize the sensitive data in the initial user behavior data, or encrypt and desensitize the sensitive data in the user behavior data when storing the user behavior data to protect user privacy.

[0067] S12, analyze user behavior data to obtain key influencing factors.

[0068] After obtaining user behavior data, user behavior analysis is performed on the user behavior data to obtain key influencing factors. Key influencing factors are used to characterize the key factors that affect user behavior.

[0069] For example, in the embodiment of the present application, a data lineage analysis method can be used to perform user behavior analysis on user behavior data to obtain key influencing factors. The key influencing factors can be understood as the analysis results of the data lineage analysis.

[0070] Data lineage analysis involves tracing and analyzing the entire process of data from its generation to its final use. The goal is to understand the data's origins, transformations, and how it is applied across various systems and business processes. For example, consider a user who frequently searches for information about gold investments during a specific time period and subsequently invests in gold during a period of strong market performance. In this case, data lineage analysis would track the user's entire journey from gold information search to actual gold investment.

[0071] Analyzing user behavior data using data lineage analysis methods to obtain key influencing factors can include the following steps:

[0072] S121: Analyze the correlation between user behavior data based on the user behavior data.

[0073] The embodiment of the present application can analyze the correlation between user behavior data based on user behavior data to restore the process of user behavior data from its generation to final use.

[0074] Association rule mining is the process of discovering frequently occurring, associated data items in a database. Therefore, embodiments of the present application can analyze the associations between user behavior data based on association rule mining algorithms, such as the Apriori algorithm and the FP-growth algorithm.

[0075] The formulas that may be used in the association rule mining algorithm are as follows:

[0076] Support refers to the frequency with which a behavior combination appears in all behaviors. It can be understood as a measure of how often a set of items (such as a product portfolio) appears in all transactions. High support indicates that a set of items often appears together and is a candidate for association rule mining. An item set can include multiple items. The formula for calculating support is:

[0077] Support(I)=Numberoftransactions containing I / Total numberoftransactions.

[0078] Among them, I represents an itemset, Number of transactions containing I is the number of transactions containing itemset I, and Total number of transactions is the total number of all transactions in the dataset.

[0079] Confidence refers to the probability of behavior B occurring under the condition that behavior A occurs. The formula for calculating confidence is:

[0080] Confidence(A→B)=Support(A∪B) / Support(A).

[0081] S122: Identify user behavior patterns based on the correlation between user behavior data to obtain key influencing factors.

[0082] Embodiments of the present application can identify user behavior patterns based on correlations between user behavior data to determine key influencing factors. For example, data mining techniques such as cluster analysis algorithms can be used to identify user behavior patterns. For example, identifying a user's behavior pattern of stockpiling gold products after viewing news about the outbreak of war indicates that the user has a high sensitivity to market dynamics. This key influencing factor is the user's high sensitivity to market dynamics. Cluster analysis algorithms such as K-means are used to group users according to their behavior patterns.

[0083] In the embodiments of this application, data lineage analysis is the key to understanding the deep-seated motivations and factors behind user behavior. By combining user behavior data with the context and influencing factors of user behavior data, users' needs and preferences can be more accurately identified and understood, thereby providing more precise and personalized services. For example, if a user group is found to frequently consult online customer service after browsing information about housing loans, this user behavior pattern may indicate that this user group is interested in buying a house but needs more guidance. By identifying this user behavior pattern, more customized information about buying a house and loans can be provided to this user group.

[0084] S13, performing similarity calculation based on the user behavior data and key influencing factors to obtain a similarity calculation result, which represents the similarity between users.

[0085] After obtaining user behavior data and key influencing factors, similarity calculation is performed to obtain similarity calculation results used to characterize the similarity between users. For example, statistical methods can be used to measure the similarity of different users' behavior in using the application.

[0086] The present application embodiments can combine user behavior data with the results of data lineage analysis through data fusion to consider the underlying factors behind user behavior. Weighted processing can be performed, giving higher weights to important factors identified by data lineage analysis, such as user reactions to specific events, to identify the key factors influencing user behavior.

[0087] Illustratively, the embodiment of the present application can obtain the similarity calculation result by calculating the Pearson correlation coefficient.

[0088] The basic formula for the Pearson correlation coefficient is as follows:

[0089]

[0090] Among them, r represents the Pearson correlation coefficient, X i represents the rating or usage frequency of user X on the i-th service, Y i represents the rating or usage frequency of user Y on the i-th service, represents the average rating or average usage frequency of user X, represents the average rating or average usage frequency of user Y, and i represents a service item in the application. For example, i=1 can represent the account management service, i=2 can represent the investment consulting service, and so on. The user's rating or usage frequency for each service can be obtained from the user behavior data.

[0091] The basic formula of the Pearson correlation coefficient is a measure of the linear relationship between two variables. It does not directly include the underlying factors. Therefore, in the embodiment of the present application, the weighted Pearson correlation coefficient is used to consider the underlying factors behind user behavior.

[0092] When calculating the Pearson correlation coefficient, different weights can be assigned to different user behavior data based on the results of data lineage analysis. For example, if the results of data lineage analysis indicate that user reactions to specific events (such as market news) are a key influencing factor, then user behavior data related to these specific events can be given a higher weight.

[0093] Therefore, in the embodiment of the present application, similarity calculation is performed based on user behavior data and key influencing factors to obtain a similarity calculation result, which may include the following steps:

[0094] S131, determining weight coefficients for various services based on key influencing factors, where services are determined based on user behavior data;

[0095] S132, calculating the weighted average value of each user's service and the weighted value of each service based on the user behavior data and the weight coefficient of each service;

[0096] S133 , performing similarity calculation based on the service weighted average value of each user and the service weighted value of each service to obtain a similarity calculation result.

[0097] The calculation formula of the weighted Pearson correlation coefficient is as follows:

[0098]

[0099] Among them, r w represents the weighted Pearson correlation coefficient, w i represents the weight coefficient of the i-th service, w iX i represents the service weight value of user X on the i-th service, represents the weighted average value of the service of user X, w i Y i represents the service weight value of user Y on the i-th service, Represents the weighted average value of services for user Y. The service weight is the weight coefficient multiplied by the rating or usage frequency of the service.

[0100] The weighted Pearson correlation coefficient obtained by the calculation formula of the weighted Pearson correlation coefficient is the similarity calculation result. In the embodiment of the present application, based on the similarity calculation results, users can be divided into different groups, each group has similar service preferences and behavior patterns, so that user groups with similar needs and preferences can be identified more accurately, thereby providing more effective personalized services. For example, user A and user B both frequently use the investment consulting service in the application, but user A prefers short-term transactions, while user B focuses more on long-term investment. By calculating the weighted Pearson correlation coefficient of these two users on various services, the similarity of user A and user B in service preferences can be evaluated.

[0101] S14, based on the similarity calculation results and key influencing factors, a recommendation algorithm is used to generate content recommendation results.

[0102] After the above process, the embodiment of the present application can obtain similarity calculation results and key influencing factors. Then, a recommendation algorithm can be used to generate content recommendation results based on the similarity calculation results and key influencing factors. For example, it can recommend products, services, or distribution rights to users. The content recommendation results may include one or more recommended content.

[0103] The embodiment of the present application can combine key influencing factors and similarity calculation results to obtain a comprehensive behavior profile of each user. The comprehensive behavior profile can include the user's preferences, behavior patterns, and potential needs.

[0104] Recommendation algorithms are used to analyze users' comprehensive behavioral profiles to generate personalized matching product or service marketing strategies based on the similarity of user behavior and personal preferences.

[0105] Recommendation algorithms may include collaborative filtering algorithms and / or content recommendation algorithms. A collaborative filtering algorithm is used to make recommendations based on similarity calculation results, resulting in a first recommendation result. A content recommendation algorithm is used to make recommendations based on user behavior and preferences, resulting in a second recommendation result. The first recommendation result may include one or more recommended content, and the second recommendation result may include one or more recommended content. The recommended content in the first recommendation result and the second recommendation result may be the same or different.

[0106] After obtaining the first and / or second recommendation results through the recommendation algorithm, embodiments of the present application can perform weighted ranking of different recommended content based on the user's behavior and preferences. This can be understood as ranking different recommended content based on key influencing factors. The order of recommended content can be adjusted based on key influencing factors. For example, for a user interested in investment, if the recommended content includes an investment product that matches their preferences, then the investment product will be ranked first in the recommended content after ranking.

[0107] The embodiments of this application not only provide recommended content that meets the user's current needs, but also predict and meet the user's future needs, making content recommendations more personalized and accurate. For example, user E often browses economic news in the information section and then checks for financial fund-related services. The results of data lineage analysis reveal user E's high sensitivity to market dynamics, and similarity calculations indicate that user E's investment behavior is highly similar to that of a certain user group. Combining this information, user E can be given upcoming financial products and market dynamic analysis reports.

[0108] In an embodiment of the present application, user behavior data is obtained; the user behavior data is analyzed to obtain key influencing factors; similarity calculation is performed based on the user behavior data and the key influencing factors to obtain a similarity calculation result, which represents the similarity between users; and a recommendation algorithm is used to generate content recommendation results based on the similarity calculation result and the key influencing factors. By combining user behavior data and key influencing factors, the embodiment of the present application comprehensively considers the background and deep-seated factors of user behavior, and can more accurately understand the user's real needs and preferences. By analyzing user behavior data, it can respond to changes in personal needs in a timely manner, making the recommended content more flexible and dynamic.

[0109] See Figure 2 , shows a flowchart of another content recommendation method provided by an embodiment of the present application, which includes at least the following steps:

[0110] S21, obtain user behavior data.

[0111] S22, analyze user behavior data to obtain key influencing factors.

[0112] S23, performing similarity calculation based on the user behavior data and key influencing factors to obtain a similarity calculation result, which represents the similarity between users.

[0113] S24, based on the similarity calculation results and key influencing factors, a recommendation algorithm is used to generate content recommendation results.

[0114] S21 to S24 can refer to Figure 1S11 to S14 in the corresponding embodiment can be understood and will not be repeated here.

[0115] S25, visually display the content recommendation results.

[0116] S26: Obtain feedback data for the content recommendation result.

[0117] S27, optimize the recommendation algorithm based on the feedback data.

[0118] The embodiments of the present application can collect user feedback data, such as through interaction tracking, collecting user interaction data on content recommendation results, such as click-through rate, browsing time, transaction conversion, etc., or collecting user direct feedback on content recommendation results through questionnaires, rating systems, or direct comments.

[0119] Use data analysis techniques, such as sentiment analysis and text analysis, to analyze user feedback and identify user acceptance of content recommendations. Collected feedback data is used to adjust and optimize recommendation algorithms, data lineage analysis methods, and / or similarity calculation methods. This can include changing recommendation weights, changing service weight coefficients, introducing new user preference parameters, and optimizing user clustering. By using user feedback data as a continuous input and continuously iterating to improve recommendation algorithms, we can capture and adapt to changes in user behavior and preferences in real time, making subsequent online marketing operations more flexible and dynamic.

[0120] Next, a content recommendation device provided in an embodiment of the present application is introduced. The content recommendation device introduced below and the content recommendation method introduced above can refer to each other.

[0121] See Figure 3 , which shows a schematic structural diagram of a content recommendation device provided in an embodiment of the present application, the device comprising:

[0122] Acquisition module 301, used to acquire user behavior data;

[0123] An analysis module 302 is used to analyze the user behavior data to obtain key influencing factors;

[0124] A calculation module 303 is configured to perform similarity calculation based on the user behavior data and the key influencing factors to obtain a similarity calculation result, wherein the similarity calculation result represents the similarity between users;

[0125] The generating module 304 is configured to generate a content recommendation result using a recommendation algorithm based on the similarity calculation result and the key influencing factors.

[0126] In the embodiment of the present application, the analysis module 302 is specifically configured to:

[0127] The user behavior data is analyzed using a data lineage analysis method to obtain the key influencing factors.

[0128] In this embodiment of the present application, the analysis module 302 includes:

[0129] an association analysis unit, configured to analyze associations between the user behavior data based on the user behavior data;

[0130] The pattern recognition unit is used to recognize the user behavior pattern based on the correlation between the user behavior data to obtain the key influencing factors.

[0131] In the embodiment of the present application, the calculation module 303 includes:

[0132] a weight determination unit, configured to determine a weight coefficient of each service based on the key influencing factors, wherein the service is determined based on the user behavior data;

[0133] A weighted calculation unit, configured to calculate a weighted average value of services for each user and a weighted value of each service based on the user behavior data and the weight coefficients of the services;

[0134] The calculation unit is configured to perform similarity calculation based on the weighted average value of the services of each user and the weighted value of the services of each service to obtain a similarity calculation result.

[0135] In an embodiment of the present application, the device further includes:

[0136] A data acquisition module, configured to acquire feedback data regarding the content recommendation results;

[0137] An optimization module is used to optimize the recommendation algorithm according to the feedback data.

[0138] The embodiment of the present application further provides a computer device, comprising: a memory and a processor;

[0139] Wherein, the memory is used to store computer programs;

[0140] The processor is configured to execute the computer program in the memory to implement the method described in the above method embodiment.

[0141] An embodiment of the present application further provides a computer-readable storage medium storing instructions, which, when executed on a computer, enables the computer to execute the method described in the above method embodiment.

[0142] In an embodiment of the present application, an acquisition module acquires user behavior data; an analysis module analyzes the user behavior data to obtain key influencing factors; a calculation module performs similarity calculation based on the user behavior data and the key influencing factors to obtain a similarity calculation result, which represents the similarity between users; and a generation module generates content recommendation results using a recommendation algorithm based on the similarity calculation result and the key influencing factors. By analyzing user behavior data, an embodiment of the present application can respond to changes in individual needs in a timely manner, making the recommended content more flexible and dynamic. By combining user behavior data and key influencing factors, it comprehensively considers the background and underlying factors of user behavior, and can more accurately understand the user's true needs and preferences.

[0143] It should be noted that the same or similar parts between the various embodiments can be referred to each other. As for the device and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0144] For the sake of simplicity, the aforementioned embodiments are described as a series of action combinations. However, those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0145] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0146] The above description of the disclosed embodiments will enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0147] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A content recommendation method, characterized in that: The method comprises: Obtain user behavior data; Analyze the user behavior data to obtain key influencing factors; Performing similarity calculation based on the user behavior data and the key influencing factors to obtain a similarity calculation result, wherein the similarity calculation result represents the similarity between users; Based on the similarity calculation result and the key influencing factors, a recommendation algorithm is used to generate a content recommendation result.

2. The method according to claim 1, characterized in that The analyzing of the user behavior data to obtain key influencing factors includes: The user behavior data is analyzed using a data lineage analysis method to obtain the key influencing factors.

3. The method according to claim 2, characterized in that The method of using data lineage analysis to analyze the user behavior data to obtain the key influencing factors includes: Analyzing the correlation between the user behavior data based on the user behavior data; The user behavior pattern is identified based on the correlation between the user behavior data to obtain the key influencing factors.

4. The method according to claim 1, characterized in that: The similarity calculation is performed based on the user behavior data and the key influencing factors to obtain a similarity calculation result, including: Determine a weight coefficient of each service based on the key influencing factors, wherein the service is determined based on the user behavior data; Calculate the weighted average value of services for each user and the weighted value of services for each service based on the user behavior data and the weight coefficients of the services; A similarity calculation is performed based on the service weighted average value of each user and the service weighted value of each service to obtain a similarity calculation result.

5. The method according to any one of claims 1 to 4, characterized in that: After generating a content recommendation result using a recommendation algorithm based on the similarity calculation result and the key influencing factors, the method further includes: Obtaining feedback data for the content recommendation result; The recommendation algorithm is optimized according to the feedback data.

6. A content recommendation device, characterized in that: The device comprises: Acquisition module, used to obtain user behavior data; An analysis module, used to analyze the user behavior data to obtain key influencing factors; A calculation module, configured to perform similarity calculation based on the user behavior data and the key influencing factors to obtain a similarity calculation result, wherein the similarity calculation result represents the similarity between users; A generation module is used to generate content recommendation results using a recommendation algorithm based on the similarity calculation result and the key influencing factors.

7. The device according to claim 6, characterized in that The analysis module is specifically used for: The user behavior data is analyzed using a data lineage analysis method to obtain the key influencing factors.

8. The device according to claim 6, characterized in that The computing module comprises: A weight determination unit, configured to determine a weight coefficient of each service based on the key influencing factors, wherein the service is determined based on the user behavior data; A weighted calculation unit, used to calculate the weighted average value of services for each user and the service weighted value of each service according to the user behavior data and the weight coefficients of each service; The calculation unit is used to perform similarity calculation based on the weighted average value of the services of each user and the weighted value of the services of each service to obtain a similarity calculation result.

9. A computer device, characterized in that: include: Memory and processor; The memory is used to store computer programs; The processor is configured to execute the computer program in the memory to implement the method according to any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that: The device stores instructions which, when executed on a computer, cause the computer to execute the method according to any one of claims 1 to 5.

Images