Information recommendation method and device, storage medium, and electronic device
By generating an item co-occurrence matrix and combining it with consumption analysis based on user profiles, the limitation of personalized recommendation algorithms in terms of recommendation scope is solved, resulting in more accurate item recommendations and a better user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HAIER YOUJIA INTELLIGENT TECH (BEIJING) CO LTD
- Filing Date
- 2022-08-30
- Publication Date
- 2026-05-29
AI Technical Summary
Existing personalized recommendation algorithms are prone to recommending incorrect content and cannot break out of the user's historical browsing interests, resulting in a limited scope of recommendations and affecting the user's browsing experience.
By analyzing the historical browsing information of multiple different user groups and the target user, an item co-occurrence matrix is generated. Combined with the user profile and consumption analysis results of the target user, a fitting process is performed to determine the distribution of the target user's interest in the items to be recommended, thereby expanding the recommendation scope.
It improves recommendation accuracy, makes recommendation results more closely resemble group characteristics, expands the scope of recommendations, and enhances the user browsing experience.
Smart Images
Figure CN115408610B_ABST
Abstract
Description
Technical Field
[0001] This application relates to personalized information recommendation technology, and more particularly to an information recommendation method, apparatus, storage medium, and electronic device. Background Technology
[0002] Personalized recommendations are a method that recommends information or items to users based on their interests and behavioral characteristics, helping them quickly discover the items they truly need from a sea of information.
[0003] Currently, commonly used personalized recommendation algorithms frequently recommend incorrect content. Specifically, these algorithms include two modes: user-based and item-based. Because user-based recommendations suffer from the problem of a large user base, item-based recommendations are often used to suggest items to users. However, item-based recommendations tend to amplify low-probability user behaviors, leading to incorrect recommendations. Furthermore, current personalized recommendation algorithms consistently fail to consider a user's historical browsing interests, thus limiting the scope of recommendations.
[0004] Therefore, how to expand the scope of recommendations and improve recommendation accuracy to enhance the user browsing experience remains a problem that needs to be solved. Summary of the Invention
[0005] This application provides an information recommendation method, apparatus, storage medium, and electronic device to address the problem of how to expand the scope of recommendations and improve recommendation accuracy in order to enhance the user browsing experience.
[0006] On the one hand, this application provides an information recommendation method, including:
[0007] The distribution of the target user's interest in the items to be recommended is determined based on the historical browsing information of multiple different user groups and the historical browsing information of the target user. The items to be recommended include multiple items.
[0008] Obtain a user profile of the target user, and obtain the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user. The consumption analysis results include at least the consumption category analysis results and the consumption level analysis results.
[0009] The distribution results are fitted based on the consumption analysis results of the user group to which they belong, to obtain the item recommendation results for the target user.
[0010] In one embodiment, determining the distribution of the target user's interest in the items to be recommended based on the historical browsing information of multiple different user groups and the target user's historical browsing information includes:
[0011] Generate an item co-occurrence matrix based on the historical browsing information of multiple different user groups;
[0012] The distribution of the target user's interest in the items to be recommended is determined based on the target user's historical browsing information and the item co-occurrence matrix.
[0013] In one embodiment, determining the distribution of the target user's interest in multiple items based on the target user's historical browsing information and the item co-occurrence matrix includes:
[0014] The initial distribution of the target user's interest in multiple items is determined based on the target user's historical browsing information and the item co-occurrence matrix.
[0015] Based on the historical browsing information of multiple different user groups, correlation analysis is performed between different categories of items. Based on the results of the correlation analysis, the distribution with correlation in the initial distribution results is selected to obtain the distribution results.
[0016] In one embodiment, determining the distribution of the target user's interest in the items to be recommended based on the target user's historical browsing information and the item co-occurrence matrix includes:
[0017] Generate an interest matrix of the target user for items viewed in the past based on the target user's browsing history;
[0018] The distribution of target users' interest in the items to be recommended is determined based on the interest matrix and the item co-occurrence matrix.
[0019] In one embodiment, determining the distribution of the target user's interest in the items to be recommended based on the interest matrix and the item co-occurrence matrix includes:
[0020] The interest level of the target user towards the items to be recommended is determined based on the interest matrix and the item co-occurrence matrix.
[0021] Based on the sorting rules and the target users' interest in the items to be recommended, the interest in the items to be recommended is filtered out. The distribution of the target users' interest in the items to be recommended is the distribution of the target users' interest in the items to be recommended.
[0022] In one embodiment, obtaining the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user includes:
[0023] Determine the user group to which the target user belongs based on the user profile of the target user;
[0024] The consumption analysis results of the user group to which the user group belongs are obtained by analyzing the user group to which the user group belongs.
[0025] In one embodiment, the consumption analysis results of the user group follow a power-law distribution.
[0026] On the other hand, this application provides an information recommendation device, comprising:
[0027] The processing module is used to determine the distribution of the target user's interest in the items to be recommended based on the historical browsing information of multiple different user groups and the historical browsing information of the target user. The items to be recommended include multiple items.
[0028] The acquisition module is used to acquire the user profile of the target user, and to acquire the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user. The consumption analysis results include at least the consumption category analysis results and the consumption level analysis results.
[0029] The processing module is also used to fit the distribution results based on the consumption analysis results of the user group to which the distribution belongs, so as to obtain the item recommendation results for the target user.
[0030] On the other hand, this application provides a computer-readable storage medium storing computer-executable instructions that, when executed, cause a computer to perform the information recommendation method as described in the first aspect.
[0031] On the other hand, this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to execute the information recommendation method as described in the first aspect through the computer program.
[0032] On the other hand, this application provides a computer program product, including a computer program that, when executed, implements the information recommendation method as described in the first aspect.
[0033] The information recommendation method provided in this application first determines the distribution of the target user's interest in the items to be recommended based on the historical browsing information of multiple different user groups and the target user's historical browsing information. Then, it determines the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user. Finally, it fits the distribution of the target user's interest in the items to be recommended based on the consumption analysis results of the user group to which the target user belongs, and obtains the final item recommendation result.
[0034] When recommending items to target users, combining the consumption analysis results of the user group to which the target user belongs makes the item recommendations more closely aligned with group characteristics, with a wider recommendation range and higher recommendation accuracy. This not only solves the problem of limited and singular recommended content, but also improves the user browsing experience. Attached Figure Description
[0035] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.
[0036] Figure 1 A schematic diagram illustrating an application scenario of the information recommendation method provided in this application;
[0037] Figure 2 A flowchart illustrating an information recommendation method provided in one embodiment of this application;
[0038] Figure 3 A schematic diagram of a power-law distribution provided for one embodiment of this application;
[0039] Figure 4 A schematic diagram of an information recommendation device provided in one embodiment of this application;
[0040] Figure 5 A schematic diagram of an electronic device provided for one embodiment of this application.
[0041] The accompanying drawings have illustrated specific embodiments of this disclosure, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concepts of this disclosure to those skilled in the art through reference to particular embodiments. Detailed Implementation
[0042] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.
[0043] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.
[0044] First, let me explain the terms used in this application:
[0045] User profile: A user profile is a visual representation of data associated with a user; in short, it's the tagging of user information.
[0046] Personalized recommendations are a method of recommending information or items to users based on their interests and behavioral characteristics, helping them quickly discover what they truly need from a sea of information. For example, when a user browses an interface, personalized recommendations of items or interfaces that the user might need are needed to increase click-through rates and pageview rates.
[0047] Currently, commonly used personalized recommendation algorithms frequently recommend incorrect content. Specifically, these algorithms include two modes: user-based and item-based. Because user-based recommendations suffer from the problem of a large user base, item-based recommendations are often used to suggest items to users. However, item-based recommendations tend to amplify low-probability user behaviors, leading to errors. Furthermore, current personalized recommendation algorithms consistently fail to consider users' historical browsing interests, limiting the scope of recommendations. If human intervention is used, human experience can cause the recommendations to lose their novelty-inducing quality.
[0048] Another recommendation method intervenes in the algorithmic results using user tags and product tags. This involves establishing user historical behavior tags to limit the recommendation results, and simultaneously using product tag similarity calculations to score and rank the recommendations, thus avoiding abrupt recommendations. However, this method still cannot escape the user's historical browsing interests, thus limiting the scope of recommendations.
[0049] Therefore, how to expand the scope of recommendations and improve recommendation accuracy to enhance the user browsing experience remains a problem that needs to be solved.
[0050] Based on this, this application provides an information recommendation method, apparatus, storage medium, and electronic device. The information recommendation method includes determining the distribution of a target user's interest in items to be recommended based on historical browsing information from multiple different user groups and the target user's historical browsing information; obtaining a user profile of the target user; obtaining consumption analysis results of the user group to which the target user belongs based on the user profile; the consumption analysis results include at least consumption category analysis results and consumption level analysis results; and fitting the distribution results to the consumption analysis results of the user group to obtain the item recommendation results for the target user.
[0051] This information recommendation method can combine the group characteristics of the target user's group to recommend items, making the item recommendation results more closely related to group characteristics, with a wider recommendation range and higher recommendation accuracy. It not only solves the problem of single and limited recommended content, but also improves the user browsing experience.
[0052] The information recommendation method provided in this application is applied to electronic devices, such as computers, servers used in laboratories, etc. Figure 1 This diagram illustrates the application of the information recommendation method. The electronic device acquires historical browsing information from multiple user groups and the target user. Based on this information, it determines the distribution of the target user's interest in the items to be recommended. Next, it acquires a user profile of the target user, determines the user group to which the target user belongs, and fits the distribution of the target user's interest in the items to be recommended based on the consumption analysis results of that user group, thus obtaining the item recommendation result for the target user.
[0053] Please see Figure 2 One embodiment of this application provides an information recommendation method, including:
[0054] S210, determine the distribution of the target user's interest in the items to be recommended based on the historical browsing information of multiple different user groups and the historical browsing information of the target user. The items to be recommended include multiple items.
[0055] The browsing history information of multiple different user groups includes a massive record of items viewed, including purchase records. Based on this browsing history information, an item co-occurrence matrix can be generated. This matrix includes both previously viewed items and items related to those items; these related items are the proposed recommended items. The multiple different user groups can include all users or a sufficiently large number of users.
[0056] The target user's browsing history includes their purchase records of browsed items, which can reflect their level of interest in those items. Based on the target user's browsing history and an item co-occurrence matrix generated from browsing history data of multiple different user groups, the distribution of the target user's interest in the recommended items can be calculated.
[0057] The following example illustrates how to determine the distribution of target users' interest in recommended items, using historical browsing data for items A and B, and proposed recommended items C and D.
[0058] If a user E has viewed both item A and item B, then to predict user E's interest in item C and item D, we can first generate an item co-occurrence matrix based on the historical browsing information (viewing item A and item B) of multiple different user groups.
[0059] The generated item co-occurrence matrix is, for example,
[0060] Then, a user interest matrix is generated based on the historical browsing information of multiple different user groups. For example, this user interest matrix might be...
[0061] Predict user E's interest in items C and D. From the interest level S, we can know that user E has an interest level of 2 for item C and an interest level of 4 for item D, indicating that user E is more interested in item D.
[0062] Based on this principle, an item co-occurrence matrix is generated using the historical browsing information of multiple different user groups. Then, the distribution of the target user's interest in the items to be recommended is determined based on the target user's historical browsing information and the item co-occurrence matrix. Specifically, first, an interest matrix of the target user on previously browsed items is generated based on the target user's historical browsing information. Then, the distribution of the target user's interest in the items to be recommended is determined based on this interest matrix and the item co-occurrence matrix.
[0063] This item co-occurrence matrix is generated based on the historical browsing information of multiple users from multiple different user groups. The amount of information is vast, often exhibiting millions or even tens of millions of dimensions. Similarly, the historical browsing information of the target users also has a relatively large number of dimensions. Therefore, the final distribution of the target users' interest in the items to be recommended is characterized by locally dense information but overall sparseness, resulting in inconsistent recommendation results.
[0064] For example, the co-occurrence matrix of items is
[0065] The target user's interest matrix for historically viewed items is as follows:
[0066] The distribution of the target user's interest in the items to be recommended.
[0067] As can be seen, although the distribution results can reflect the target user's interest in the recommended items, the difference in interest between different items is large, and further processing of the distribution results is needed.
[0068] In an optional embodiment, based on the sorting rules and the target user's interest in the items to be recommended, the interest levels of the items to be recommended are filtered out from the list of items to be recommended. The distribution of the target user's interest in the items to be recommended is the distribution of the target user's interest in the items to be recommended. For example, the target user's interest in the items to be recommended is S = [5 0 0 0 0 0 1.6 0 0.1] as described above, and the interest levels of the items to be recommended in the filtered list are 5, 1.6, and 0.1.
[0069] In an optional embodiment, an initial distribution of the target user's interest in multiple items is first determined based on the target user's historical browsing information and the item co-occurrence matrix. This initial distribution could be, for example, [5 0 0 00 0 1.6 0 0.1] as described above, or the interest level corresponding to the item to be recommended as described above.
[0070] Then, based on the historical browsing information of these multiple user groups, correlation analysis is performed between different categories of items. Based on the results of the correlation analysis, the distributions with correlation in the initial distribution results are selected to obtain the final distribution. Selecting the distributions with correlation in the initial distribution results based on the correlation analysis results is a reasonable filter of the initial distribution results, that is, filtering out the interest distribution results of some items with strong correlation as the desired interest distribution results.
[0071] Association analysis identifies related behaviors by analyzing actual human actions. Metrics used in association analysis include support, association strength, and lift. Support refers to the probability that a set (e.g., a specific item) appears in the overall set (e.g., all items). Confidence refers to the probability that condition Y is derived from association rule XY given that precondition X has occurred. Lift is the ratio of the probability that condition Y is also present given precondition X to the overall probability of condition Y occurring. It's easy to see that the support is higher when two items frequently co-occur. Confidence verifies whether two items are associated, while lift quantifies the correlation between them.
[0072] Finally, based on the results of association analysis, the distribution with correlation in the initial distribution results is selected as the distribution of the target user's interest in the items to be recommended.
[0073] S220, Obtain the user profile of the target user, and obtain the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user. The consumption analysis results include at least the consumption category analysis results and the consumption level analysis results.
[0074] As the user base expands, individual user characteristics gradually converge into group characteristics. Recommending items based on group characteristics not only addresses the issues of limited and singular content but also improves user experience. Therefore, incorporating the characteristics of the target user's user group into item recommendations can enhance recommendation effectiveness.
[0075] User profiling is used to tag user information. A target user's profile can be determined based on their historical spending records, and then the user group to which that target user belongs can be identified based on that profile. Analysis of this user group yields consumption analysis results. Specifically, analyzing the consumption characteristics of this user group produces consumption category analysis results and consumption level analysis results. The consumption category analysis results may include, for example, the types of goods the user group prefers to consume, and the consumption level analysis results may include, for example, the levels of goods the user group prefers to consume.
[0076] S230: Based on the consumption analysis results of the user group to which it belongs, the distribution results are fitted to obtain the item recommendation results for the target user.
[0077] The consumption analysis results of this user group follow a power-law distribution. The power-law distribution, also known as the Matthew effect, is a statistical concept. The probability density function of a power-law distribution can be expressed as y = αx. -γ The form is , where x and y are positive random variables, and α and γ are both positive constants. In a power-law distribution, higher probabilities account for a smaller percentage, and a large percentage are those with low probabilities.
[0078] Please see Figure 3 , Figure 3 The diagram shows the power-law distribution y = 5x -2 The diagram illustrates the power function, with the tail (boxed portion) representing the high-probability but low-percentage segment. When recommending novelty items, it's necessary to suggest items that generate high user interest but represent a small percentage of spending, ensuring the recommended items are both novel and aligned with user consumption habits.
[0079] When fitting the distribution results based on the consumption analysis of the user group to which the target user belongs, the metrics used in the association analysis of the initial distribution results, as described above, including support, correlation, and lift, can be adjusted to ensure that the category and characteristic distribution of the items to be recommended in the final distribution results are consistent with the power-law distribution of the user group. Based on this consistency with the power-law distribution, items with high interest but low consumption proportions in the power-law distribution can be preferentially selected to obtain the item recommendation results for the target user. The resulting item recommendation structure for the target user accurately reflects the interest in the recommended items while also considering the group characteristics of the user group to which the target user belongs. This avoids recommending unpopular items, improves recommendation accuracy, expands the recommendation scope, and ultimately enhances the user browsing experience.
[0080] In summary, this embodiment provides an information recommendation method, including: determining the distribution of a target user's interest in items to be recommended based on the historical browsing information of multiple different user groups and the historical browsing information of a target user, wherein the items to be recommended include multiple items; obtaining a user profile of the target user, and obtaining the consumption analysis results of the user group to which the target user belongs based on the user profile, wherein the consumption analysis results include at least consumption category analysis results and consumption level analysis results; and fitting the distribution results based on the consumption analysis results of the user group to obtain the item recommendation results for the target user.
[0081] In other words, the distribution of the target user's interest in the items to be recommended is first determined based on the browsing history of multiple different user groups and the target user's browsing history. Then, the consumption analysis results of the user group to which the target user belongs are determined based on the user profile of the target user. The distribution of the target user's interest in the items to be recommended is then fitted and processed based on the consumption analysis results of the user group to which the target user belongs, resulting in the final item recommendation results.
[0082] When recommending items to target users, combining the consumption analysis results of the user group to which the target user belongs makes the item recommendations more closely aligned with group characteristics, with a wider recommendation range and higher recommendation accuracy. This not only solves the problem of limited and singular recommended content, but also improves the user browsing experience.
[0083] Please see Figure 4 An embodiment of this application also provides an information recommendation device 10, comprising:
[0084] The processing module 11 is used to determine the distribution of the target user's interest in the items to be recommended based on the historical browsing information of multiple different user groups and the historical browsing information of the target user. The items to be recommended include multiple items.
[0085] The acquisition module 12 is used to acquire the user profile of the target user and, based on the user profile of the target user, acquire the consumption analysis results of the user group to which the target user belongs. The consumption analysis results include at least the consumption category analysis results and the consumption level analysis results.
[0086] The processing module 11 is also used to fit the distribution result based on the consumption analysis results of the user group to which it belongs, so as to obtain the item recommendation result for the target user.
[0087] Specifically, the processing module 11 is used to generate an item co-occurrence matrix based on the historical browsing information of multiple different user groups; and to determine the distribution of the target user's interest in the items to be recommended based on the target user's historical browsing information and the item co-occurrence matrix.
[0088] The processing module 11 is specifically used to determine the initial distribution of the target user's interest in multiple items based on the target user's historical browsing information and the item co-occurrence matrix; to perform correlation analysis between different categories of items based on the historical browsing information of the multiple different user groups; and to filter out the distributions with correlation in the initial distribution based on the results of the correlation analysis, thereby obtaining the distribution result.
[0089] Specifically, the processing module 11 is used to generate an interest matrix of the target user on the items browsed in the past based on the target user's historical browsing information; and to determine the distribution of the target user's interest in the items to be recommended based on the interest matrix and the item co-occurrence matrix.
[0090] The processing module 11 is specifically used to determine the target user's interest in the proposed recommended items based on the interest degree matrix and the item co-occurrence matrix; according to the sorting rules and the target user's interest in the proposed recommended items, it filters out the interest degree corresponding to the items to be recommended among the proposed recommended items, and the distribution result of the target user's interest in the proposed recommended items is the distribution result of the target user's interest in the items to be recommended.
[0091] The acquisition module 12 is specifically used to determine the user group to which the target user belongs based on the user profile of the target user; analyze the user group to which the target user belongs, and obtain the consumption analysis results of the user group to which the target user belongs. The consumption analysis results of the user group to which the target user belongs follow a power-law distribution.
[0092] Please see Figure 5 An embodiment of this application also provides an electronic device 20, including a memory 21 and a processor 22, wherein the memory 21 stores a computer program and the processor 22 is configured to execute the information recommendation method provided in any of the preceding embodiments through the computer program.
[0093] This application also provides a computer-readable storage medium storing computer-executable instructions that, when executed, cause a computer to perform the information recommendation method provided in any of the above embodiments.
[0094] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the information recommendation method as provided in any of the above embodiments.
[0095] It should be noted that the aforementioned computer-readable storage media can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM), etc. It can also be various electronic devices that include one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.
[0096] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0097] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0098] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.
[0099] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0100] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0101] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0102] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. An information recommendation method, characterized in that, include: An item co-occurrence matrix is generated based on the historical browsing information of multiple different user groups. The item co-occurrence matrix includes historically browsed items and items related to historically browsed items. The distribution of the target user's interest in the items to be recommended is determined based on the target user's historical browsing information and the item co-occurrence matrix. The items to be recommended include multiple items. Obtain a user profile of the target user, and obtain the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user. The consumption analysis results include at least the consumption category analysis results and the consumption level analysis results. The distribution results are fitted based on the consumption analysis results of the user group to which they belong, to obtain the item recommendation results for the target user.
2. The method according to claim 1, characterized in that, The determination of the target user's interest distribution among multiple items based on the target user's historical browsing information and the item co-occurrence matrix includes: The initial distribution of the target user's interest in multiple items is determined based on the target user's historical browsing information and the item co-occurrence matrix. Based on the historical browsing information of multiple different user groups, correlation analysis is performed between different categories of items. Based on the results of the correlation analysis, the distribution with correlation in the initial distribution results is selected to obtain the distribution results.
3. The method according to claim 1, characterized in that, The determination of the target user's interest distribution in the items to be recommended based on the target user's historical browsing information and the item co-occurrence matrix includes: Generate an interest matrix of the target user for items viewed in the past based on the target user's browsing history; The distribution of target users' interest in the items to be recommended is determined based on the interest matrix and the item co-occurrence matrix.
4. The method according to claim 3, characterized in that, The determination of the distribution of target users' interest in the items to be recommended based on the interest matrix and the item co-occurrence matrix includes: The interest level of the target user towards the items to be recommended is determined based on the interest matrix and the item co-occurrence matrix. Based on the sorting rules and the target users' interest in the items to be recommended, the interest in the items to be recommended is filtered out. The distribution of the target users' interest in the items to be recommended is the distribution of the target users' interest in the items to be recommended.
5. The method according to any one of claims 1-4, characterized in that, The step of obtaining the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user includes: Determine the user group to which the target user belongs based on the user profile of the target user; The consumption analysis results of the user group to which the user group belongs are obtained by analyzing the user group to which the user group belongs.
6. The method according to claim 5, characterized in that, The consumption analysis results of the user group to which the data belongs follow a power-law distribution.
7. An information recommendation device, characterized in that, include: The processing module is used to generate an item co-occurrence matrix based on the historical browsing information of multiple different user groups. The item co-occurrence matrix includes historically browsed items and items related to historically browsed items. Based on the target user's historical browsing information and the item co-occurrence matrix, the distribution of the target user's interest in the items to be recommended is determined. The items to be recommended include multiple items. The acquisition module is used to acquire the user profile of the target user, and to acquire the consumption analysis results of the user group to which the target user belongs based on the user profile of the target user. The consumption analysis results include at least the consumption category analysis results and the consumption level analysis results. The processing module is also used to fit the distribution results based on the consumption analysis results of the user group to which the distribution belongs, so as to obtain the item recommendation results for the target user.
8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when executed, cause the computer to perform the information recommendation method as described in any one of claims 1-6.
9. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the information recommendation method as described in any one of claims 1-6 through the computer program.