A content recommendation method and system based on local life service platform

Abstract

The present invention provides a content recommendation method and system based on a local life service platform. The method includes determining a user content score based on content interaction information, and sorting the to-be-recommended content in a to-be-recommended content set in descending order based on the user content score; determining a temporal content score based on content preference information, and sorting the to-be-recommended content in a to-be-recommended content set in descending order based on the temporal content score; determining a content search score based on content search information, and sorting the to-be-recommended content in a to-be-recommended content set in descending order based on the content search score; weightedly fusing a first sorted content set, a second sorted content set, and a third sorted content set, and recommending corresponding to-be-recommended content to corresponding target users according to the sorting relationship in the final sorted content set. The present invention can improve the matching degree between recommended content and user preferences, and improve user experience.

Description

Technical Field

[0001] The present invention belongs to the technical field of content recommendation, and specifically relates to a content recommendation method and system based on a local life service platform. Background Art

[0002] For local life service platforms, they are usually online comprehensive service platforms that can meet some of the users' life service needs, such as ordering meals, buying movie tickets, buying flowers, etc., and for existing life service platforms, in order to improve the user experience, they can usually push content that meets the user's needs to the corresponding users based on the user's preference information to enhance the user's experience. As for existing content recommendation methods, they usually simply push content based on the user's historical records, which is not comprehensive enough, resulting in the final pushed content not matching the user's preferences, affecting the user's experience. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a content recommendation method and system based on a local life service platform, which are used to solve the technical problems in the prior art.

[0004] On the one hand, the present invention provides the following technical solution, a content recommendation method based on a local life service platform, comprising:

[0005] Obtaining content interaction information of a target user on a local life service platform, determining a user content score based on the content interaction information, and sorting the to-be-recommended content in the to-be-recommended content set in descending order based on the user content score to obtain a first sorted content set;

[0006] determining content preference information based on the content interaction information, determining a time-series content score based on the content preference information, and sorting the to-be-recommended content in the to-be-recommended content set in descending order based on the time-series content score to obtain a second sorted content set;

[0007] Obtaining content search information of the target user on the local life service platform, determining a content search score based on the content search information, and sorting the to-be-recommended content in the to-be-recommended content set in descending order based on the content search score to obtain a third sorted content set;

[0008] The first sorted content set, the second sorted content set, and the third sorted content set are weightedly merged to obtain a final sorted content set, and corresponding to-be-recommended content is recommended to corresponding target users according to the sorting relationship in the final sorted content set.

[0009] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention first obtains the content interaction information of the target user on the local life service platform, determines the user content score based on the content interaction information, and sorts the content to be recommended in the content set to be recommended in descending order based on the user content score to obtain a first sorted content set; then determines the content preference information based on the content interaction information, determines the time content score based on the content preference information, and sorts the content to be recommended in the content set to be recommended in descending order based on the time content score to obtain a second sorted content set; then obtains the content search information of the target user on the local life service platform, determines the content search score based on the content search information, and sorts the content to be recommended in the content set to be recommended in descending order based on the content search score to obtain a third sorted content set; finally, the first sorted content set, the second sorted content set, and the third sorted content set are weightedly merged to obtain a final sorted content set, and the corresponding content to be recommended is recommended to the corresponding target user according to the sorting relationship in the final sorted content set. The present invention comprehensively considers the influence of user preferences, timing, and the influence of user search content on content push, and then determines three sorting groups, and then weightedly integrates the three sorting groups, so as to improve the matching degree between recommended content and user preferences and improve user experience.

[0010] Preferably, the step of determining the user content rating based on the content interaction information includes:

[0011] Extracting the number of interactions and content labels from the content interaction information, constructing a user hypergraph with users as hyperedges and content labels as hyperpoints, and constructing a content hypergraph with user interaction content as hyperedges and content labels as superpoints;

[0012] Calculate the first user weight of the user hypergraph With the second user weight :

[0013] ;

[0014] ;

[0015] In the formula, Indicates the amount of user interaction content, Indicates User interaction content, Indicates The number of interactions of the user's interactive content, express A collection of content tags. Content label The order of markings;

[0016] Calculate the first content weight of the content hypergraph With the second content weight :

[0017] ;

[0018] ;

[0019] In the formula, Indicates user The total number of interactions, Indicates the number of users, Indicates user With content tags The number of interactions;

[0020] In the user hypergraph, a hyperpoint is selected to iteratively transfer toward the adjacent hyperedge, and the first iteration vector after the iterative transfer is calculated. , select a superpoint in the content hypergraph to iteratively transfer toward the adjacent hyperedge, and calculate the second iteration vector after iterative transfer :

[0021] ;

[0022] ;

[0023] ;

[0024] ; ;

[0025] ; ;

[0026] In the formula, , Represent the number of superpoints in the user hypergraph and content hypergraph respectively, is the iteration factor, , They represent the hyperedge vectors in the user hypergraph and content hypergraph respectively. denote the first incidence matrix and the second incidence matrix respectively, Respectively The elements in They represent the first Super point, A super edge, Respectively represent the first Super point, A super edge, , , , denote the first user diagonal matrix, the second user diagonal matrix, the first content diagonal matrix, and the second content diagonal matrix, respectively. Respectively , , , The elements in , Respectively represent The first iteration vector after iterations, , Respectively represent The second iteration vector after iterations;

[0027] Repeat the iterative transfer process until the iterative condition is met, and output the corresponding first iterative vector and second iterative vector as the first final vector and the second final vector respectively;

[0028] A user content score is determined based on the first final vector and the second final vector.

[0029] Preferably, the step of determining the user content score based on the first final vector and the second final vector includes:

[0030] Calculate the user vector based on the first final vector and the second final vector With content vector :

[0031] ;

[0032] ;

[0033] In the formula, , Representation Matrix , Middle Line The elements of the column, , Representation Matrix , Middle Line The elements of the column, represents the user-label matrix, express The adjacency matrix of , Indicates the first final vector, the second final vector elements, Indicates the number of content tags;

[0034] Based on user vector With content vector Calculate user score vector and content score vector :

[0035] ; ;

[0036] In the formula, , Set the user diagonal matrix for the first and second respectively, are the first and second setting content diagonal matrices respectively;

[0037] Based on user score vector and content score vector Determining User Content Ratings :

[0038] .

[0039] Preferably, the step of determining the temporal content score based on the content preference information includes:

[0040] Extract evaluation information of the target user and neighbor users from the content preference information, and determine a first target evaluation value based on the evaluation information , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value :

[0041] ; ;

[0042] ; ;

[0043] In the formula, represents the time control factor, is the maximum timestamp, , Respectively represent the target users With neighbor users User interaction content Evaluation, Indicate whether you like or dislike the review. Indicates the target user In the user interaction content Timestamps for likes and dislikes, Indicates neighbor users In the user interaction content Timestamps for likes and dislikes;

[0044] Based on the first target evaluation value , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value Calculating user similarity :

[0045] ;

[0046] ;

[0047] In the formula, Represents user interaction content The comprehensive evaluation value of

[0048] The target user and neighbor users' preferences for the user interaction content are divided into index layers to obtain several layers of index sets, the membership degrees corresponding to the several layers of index sets are determined, and the first reliability value of the target user is calculated based on the membership degrees. The second most reliable value with neighbor users :

[0049] ;

[0050] ;

[0051] In the formula, Indicates the target user and neighbor user's interaction content The membership degree of , It represents the average membership degree of each layer indicator set corresponding to the target user and neighbor users. , Indicates the upper and lower limits of the user evaluation interval, , Indicates the target user and neighbor user's interaction content The actual evaluation value of

[0052] First reliable value based on target user The second most reliable value with neighbor users Determine temporal content rating.

[0053] Preferably, the first reliability value based on the target user The second most reliable value with neighbor users The steps to determine the temporal content rating include:

[0054] Based on the first reliable value The second reliable value Calculation time reliability :

[0055] ;

[0056] In the formula, , Represents the evaluation set of the target user and neighbor users;

[0057] Time-based reliability Determine the set of neighbor users :

[0058] ;

[0059] In the formula, represents the initial neighbor user set, is the reliability threshold;

[0060] Based on neighbor user set Determining Timed Content Ratings :

[0061] .

[0062] Preferably, the step of determining the content search score based on the content search information includes:

[0063] Extract the user's search keyword from the content search information, and calculate the similarity between the search keyword and the candidate keywords in the candidate keyword set :

[0064] ;

[0065] In the formula, Indicates The search keyword corresponds to , Tag categories, Indicates The candidate keywords correspond to , Tag categories, , express , The common nodes of , express , The depth at which represents the distance calculation, , Respectively represent the number of label categories corresponding to the search keywords and candidate keywords;

[0066] Eliminate candidate keywords whose similarity is less than a similarity threshold from the candidate keyword set to obtain an eliminated keyword set;

[0067] Identify target users For the first membership degree of keywords in the keyword set and user interaction content For the second membership degree of keywords in the keyword set ;

[0068] According to target users The fuzzy set of the eliminated keywords is converted into several first semantic vectors, and the first semantic vectors are averaged to obtain the first average vector , based on the user interaction content The fuzzy set of the eliminated keywords is converted into several second semantic vectors, and the average of several second semantic vectors is taken to obtain the second average vector ;

[0069] Based on the first membership , the second membership , the first mean vector , the second mean vector Calculate the first keyword score vector and the second keyword score vector :

[0070] ;

[0071] ;

[0072] Based on the first keyword score vector and the second keyword score vector Calculate content search score:

[0073] ;

[0074] In the formula, Represents cosine similarity.

[0075] Preferably, the step of weightedly fusing the first sorted content set, the second sorted content set, and the third sorted content set to obtain a final sorted content set includes:

[0076] assigning first weights to the to-be-recommended content in the first sorted content set in descending order according to the arrangement relationship of the first sorted content set;

[0077] assigning second weights to the to-be-recommended content in the second sorted content set in descending order according to the arrangement relationship of the second sorted content set;

[0078] assigning third weights to the to-be-recommended content in the third sorted content set in descending order according to the arrangement relationship of the third sorted content set;

[0079] Calculating the sum of the first weight, the second weight, and the third weight of the same content to be recommended in the first ranked content set, the second ranked content set, and the third ranked content set to obtain a comprehensive weight;

[0080] The contents to be recommended in the content set to be recommended are sorted in descending order according to the size of the comprehensive weights to obtain a final sorted content set.

[0081] In a second aspect, the present invention provides the following technical solution: a content recommendation system based on a local life service platform, the system comprising:

[0082] A first sorting module is used to obtain content interaction information of a target user on a local life service platform, determine a user content score based on the content interaction information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the user content score to obtain a first sorted content set;

[0083] a second sorting module, configured to determine content preference information based on the content interaction information, determine a time-series content score based on the content preference information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the time-series content score to obtain a second sorted content set;

[0084] A third sorting module is used to obtain content search information of the target user on the local life service platform, determine a content search score based on the content search information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the content search score to obtain a third sorted content set;

[0085] The final sorting module is used to perform weighted fusion on the first sorted content set, the second sorted content set, and the third sorted content set to obtain a final sorted content set, and recommend the corresponding to-be-recommended content to the corresponding target user according to the sorting relationship in the final sorted content set.

[0086] In the third aspect, the present invention provides the following technical solution: a computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the content recommendation method based on the local life service platform as described above is implemented.

[0087] In a fourth aspect, the present invention provides the following technical solution: a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the content recommendation method based on the local life service platform as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0089] Figure 1 A flowchart of a content recommendation method based on a local life service platform provided in Embodiment 1 of the present invention;

[0090] Figure 2 A structural block diagram of a content recommendation system based on a local life service platform provided in Embodiment 2 of the present invention;

[0091] Figure 3 A schematic diagram of the hardware structure of a computer provided in another embodiment of the present invention.

[0092] The embodiments of the present invention will be further described below with reference to the accompanying drawings. DETAILED DESCRIPTION

[0093] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the embodiments of the present invention, and should not be construed as limiting the present invention.

[0094] In the description of the embodiments of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0095] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0096] In the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the embodiments of the present invention can be understood according to specific circumstances.

[0097] Embodiment 1

[0098] In the first embodiment of the present invention, Figure 1 As shown, a content recommendation method based on a local life service platform includes:

[0099] S1. Obtain content interaction information of a target user on a local life service platform, determine a user content score based on the content interaction information, and sort the to-be-recommended content in a to-be-recommended content set in descending order based on the user content score to obtain a first sorted content set;

[0100] Specifically, content interaction information includes content interaction information, number of interactions, content tags, etc. The user interaction content here includes but is not limited to videos, articles, commodities, etc., and the number of interactions here can be expressed as the number of times a user views, purchases, clicks, searches, etc. for a certain content. For content such as videos and articles, it can also be expressed as the number of likes, favorites, reposts, comments, etc. The content tag is the key tag of the content, and can also be expressed as the classification of the content.

[0101] Wherein, the step S1 comprises:

[0102] S11, extracting the number of interactions and content labels in the content interaction information, constructing a user hypergraph with users as hyperedges and content labels as hyperpoints, and constructing a content hypergraph with user interaction content as hyperedges and content labels as superpoints;

[0103] Specifically, a hypergraph is a graph structure that can connect two or more points on an edge, which includes a superpoint set, a hyperedge set, and a weight set.

[0104] S12. Calculate the first user weight of the user hypergraph With the second user weight :

[0105] ;

[0106] ;

[0107] In the formula, Indicates the amount of user interaction content, Indicates User interaction content, Indicates The number of interactions of the user's interactive content, express A collection of content tags. Content label The order of markings;

[0108] Specifically, when a user interacts with user interactive content, the user can label the content accordingly. Actually it is the number of super points in the user's hypergraph.

[0109] S13: Calculate the first content weight of the content hypergraph With the second content weight :

[0110] ;

[0111] ;

[0112] In the formula, Indicates user The total number of interactions, Indicates the number of users, Indicates user With content tags The number of interactions;

[0113] Specifically, a greater number of interactions indicates that the user has a higher interest in the content corresponding to the content tag.

[0114] S14, selecting a superpoint in the user hypergraph to iteratively transfer toward the adjacent hyperedge, and calculating the first iteration vector after the iterative transfer , select a superpoint in the content hypergraph to iteratively transfer toward the adjacent hyperedge, and calculate the second iteration vector after iterative transfer :

[0115] ;

[0116] ;

[0117] ;

[0118] ; ;

[0119] ; ;

[0120] In the formula, , Represent the number of superpoints in the user hypergraph and content hypergraph respectively, is the iteration factor, , They represent the hyperedge vectors in the user hypergraph and content hypergraph respectively. denote the first incidence matrix and the second incidence matrix respectively, Respectively The elements in They represent the first Super point, A super edge, Respectively represent the first Super point, A super edge, , , , denote the first user diagonal matrix, the second user diagonal matrix, the first content diagonal matrix, and the second content diagonal matrix, respectively. Respectively , , , The elements in , Respectively represent The first iteration vector after iterations, , Respectively represent The second iteration vector after iterations;

[0121] Specifically, in the actual transfer process, a superpoint is selected as the starting point, and the starting point is transferred toward the adjacent superedge, and then a point is selected in the transferred superedge to continue to repeat the transfer process, where: , .

[0122] S15, repeating the iterative transfer process until the iterative condition is met, and outputting the corresponding first iterative vector and second iterative vector as the first final vector and the second final vector respectively;

[0123] Specifically, after a certain number of iterations, if the values ​​in the first iteration vector and the second iteration vector converge and remain unchanged, the iteration condition is considered to be met and the iteration process is stopped, and the first iteration vector and the second iteration vector after the last iteration are output as the first final vector and the second final vector, respectively.

[0124] S16, determining a user content score based on the first final vector and the second final vector;

[0125] Wherein, the step S16 comprises:

[0126] S161, calculating user vectors based on the first final vector and the second final vector With content vector :

[0127] ;

[0128] ;

[0129] In the formula, , Representation Matrix , Middle Line The elements of the column, , Representation Matrix , Middle Line The elements of the column, represents the user-label matrix, express The adjacency matrix of , Indicates the first final vector, the second final vector elements, Indicates the number of content tags;

[0130] Among them, for the user-tag matrix, its internal elements are the number of interactions between the user and the content tag.

[0131] S161, based on user vector With content vector Calculate user score vector and content score vector :

[0132] ; ;

[0133] In the formula, , Set the user diagonal matrix for the first and second respectively, are the first and second setting content diagonal matrices respectively;

[0134] Among them, for the first set user diagonal matrix, its diagonal elements are , ,……, , , for the second set of user diagonal matrices, its diagonal elements are , ,……, , , for the first set content matrix, its diagonal elements are , ,……, , , for the second setting content diagonal matrix, its diagonal elements are , ,……, , .

[0135] S161. Based on user score vector and content score vector Determining User Content Ratings :

[0136] .

[0137] S2, determining content preference information based on the content interaction information, determining a time-series content score based on the content preference information, and sorting the to-be-recommended content in the to-be-recommended content set in descending order based on the time-series content score to obtain a second sorted content set;

[0138] The content signal information specifically refers to the user's preference rating and degree of preference for a certain content.

[0139] Wherein, the step S2 comprises:

[0140] S21, extracting evaluation information of the target user and neighbor users from the content preference information, and determining a first target evaluation value according to the evaluation information , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value :

[0141] ; ;

[0142] ; ;

[0143] In the formula, represents the time control factor, is the maximum timestamp, , Respectively represent the target users With neighbor users User interaction content Evaluation, Indicates whether you like or dislike the review. Indicates the target user In the user interaction content Timestamps for likes and dislikes, Neighbor user In the user interaction content Timestamps for likes and dislikes;

[0144] The neighbor users may be selected from the initial neighbor user set, and the larger the time control factor is, the greater the influence of time on the user signal is.

[0145] S22, based on the first target evaluation value , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value Calculating user similarity :

[0146] ;

[0147] ;

[0148] In the formula, Represents user interaction content The comprehensive evaluation value of .

[0149] S23, stratifying the target user and neighbor users' preferences for the user interaction content into indicators to obtain several layers of indicator sets, determining the memberships corresponding to the several layers of indicator sets, and calculating the first reliability value of the target user based on the memberships The second most reliable value with neighbor users :

[0150] ;

[0151] ;

[0152] In the formula, Indicates the target user and neighbor user's interaction content The membership degree of , It represents the average membership degree of each layer indicator set corresponding to the target user and neighbor users. , Indicates the upper and lower limits of the user evaluation interval, , Indicates the target user and neighbor user's interaction content The actual evaluation value of

[0153] Specifically, when scoring content, the weight of each factor is quantified, so the user's preference level is divided into several levels through indicators. For example, a number between 1-5 can be used to represent the user's preference level. The degree of membership here can be determined by a fuzzy algorithm, and the fuzzy algorithm is a commonly used algorithm in the prior art, so it will not be described in detail.

[0154] S24: First reliability value based on target user The second most reliable value with neighbor users Determine the timing content rating;

[0155] Wherein, the step S24 comprises:

[0156] S241, based on the first reliable value The second reliable value Calculation time reliability :

[0157] ;

[0158] In the formula, , Represents the evaluation set of the target user and neighbor users;

[0159] Specifically, the purpose of calculating time reliability is to eliminate invalid users to prevent them from affecting the final score calculation.

[0160] S242, based on time reliability Determine the set of neighbor users :

[0161] ;

[0162] In the formula, represents the initial neighbor user set, is the reliability threshold.

[0163] S243, based on neighbor user set Determining Timed Content Ratings :

[0164] .

[0165] S3. Obtain content search information of the target user on the local life service platform, determine a content search score based on the content search information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the content search score to obtain a third sorted content set;

[0166] Wherein, the step S3 comprises:

[0167] S31: extracting the user's search keyword from the content search information, and calculating the similarity between the search keyword and the candidate keywords in the candidate keyword set :

[0168] ;

[0169] In the formula, Indicates The search keyword corresponds to , Tag categories, Indicates The candidate keywords correspond to , Tag categories, , express , The common nodes of , express , The depth at which represents the distance calculation, , Respectively represent the number of label categories corresponding to the search keywords and candidate keywords.

[0170] S32: Remove candidate keywords whose similarity is less than a similarity threshold from the candidate keyword set to obtain a removed keyword set.

[0171] S33. Determine target users For the first membership degree of keywords in the keyword set and user interaction content For the second membership degree of keywords in the keyword set ;

[0172] S34. According to target users The fuzzy set of the eliminated keywords is converted into several first semantic vectors, and the first semantic vectors are averaged to obtain the first average vector , based on the user interaction content The fuzzy set of the eliminated keywords is converted into several second semantic vectors, and the average of several second semantic vectors is taken to obtain the second average vector .

[0173] S35, based on the first membership , the second membership , the first mean vector , the second mean vector Calculate the first keyword scoring vector and the second keyword scoring vector :

[0174] ;

[0175] .

[0176] S36. Based on the first keyword scoring vector and the second keyword scoring vector calculate the content search score:

[0177] ;

[0178] wherein, represents the cosine similarity.

[0179] S4. Weightedly fuse the first sorted content set, the second sorted content set, and the third sorted content set to obtain a final sorted content set, and recommend the corresponding content to be recommended to the corresponding target user according to the sorting relationship in the final sorted content set;

[0180] wherein, the step S4 includes:

[0181] S41. Assign a first weight to the content to be recommended in the first sorted content set according to the arrangement relationship of the first sorted content set in a decreasing relationship.

[0182] S42. Assign a second weight to the content to be recommended in the second sorted content set according to the arrangement relationship of the second sorted content set in a decreasing relationship.

[0183] S43. Assign a third weight to the content to be recommended in the third sorted content set according to the arrangement relationship of the third sorted content set in a decreasing relationship.

[0184] S44. Calculate the sum of the first weight, the second weight, and the third weight of the same content to be recommended in the first sorted content set, the second sorted content set, and the third sorted content set to obtain a comprehensive weight;

[0185] Specifically, assuming that there are 4 user interaction contents in the first sorted content group, 0.4, 0.3, 0.2, and 0.1 can be assigned as the first weights respectively, and the same applies to the second weight and the third weight. At the same time, the same user interaction content here refers to the user interaction content with the same label. Then, the first to third weights of the same user interaction content are added to obtain the comprehensive weight. Then, the comprehensive weight is sorted to obtain the final sorted content set. At the same time, if the comprehensive weights are the same, the order of the user interaction contents with the same comprehensive weights is randomly determined.

[0186] S45. Sort the contents to be recommended in the content set to be recommended in descending order according to the size of the comprehensive weights to obtain a final sorted content set.

[0187] The content recommendation method based on the local life service platform provided in the first embodiment of the present invention first obtains the content interaction information of the target user on the local life service platform, determines the user content score based on the content interaction information, and sorts the to-be-recommended content in the to-be-recommended content set in descending order based on the user content score to obtain a first sorted content set; then determines the content preference information based on the content interaction information, determines the time series content score based on the content preference information, and sorts the to-be-recommended content in the to-be-recommended content set in descending order based on the time series content score to obtain a second sorted content set; then obtains the content search information of the target user on the local life service platform, determines the content search score based on the content search information, and sorts the to-be-recommended content in the to-be-recommended content set in descending order based on the content search score to obtain a third sorted content set; finally, the first sorted content set, the second sorted content set, and the third sorted content set are weightedly merged to obtain a final sorted content set, and the corresponding to-be-recommended content is recommended to the corresponding target user according to the sorting relationship in the final sorted content set. The present invention comprehensively considers the influence of user preferences, timing, and the influence of user search content on content push, and then determines three sorting groups, and then weightedly integrates the three sorting groups, so as to improve the matching degree between recommended content and user preferences and improve user experience.

[0188] Embodiment 2

[0189] like Figure 2 As shown, in the second embodiment of the present invention, a content recommendation system based on a local life service platform is provided, and the system includes:

[0190] A first sorting module 1 is used to obtain content interaction information of a target user on a local life service platform, determine a user content score based on the content interaction information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the user content score to obtain a first sorted content set;

[0191] A second sorting module 2 is configured to determine content preference information based on the content interaction information, determine a time-series content score based on the content preference information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the time-series content score to obtain a second sorted content set;

[0192] A third sorting module 3 is used to obtain content search information of the target user on the local life service platform, determine a content search score based on the content search information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the content search score to obtain a third sorted content set;

[0193] A final sorting module 4 is used to perform weighted fusion on the first sorted content set, the second sorted content set, and the third sorted content set to obtain a final sorted content set, and recommend corresponding content to be recommended to corresponding target users according to the sorting relationship in the final sorted content set;

[0194] The first sorting module 1 comprises:

[0195] A construction submodule is used to extract the number of interactions and content labels in the content interaction information, construct a user hypergraph with users as hyperedges and content labels as hyperpoints, and construct a content hypergraph with user interaction content as hyperedges and content labels as superpoints;

[0196] A user weight submodule is used to calculate the first user weight of the user hypergraph. With the second user weight :

[0197] ;

[0198] ;

[0199] In the formula, Indicates the amount of user interaction content, Indicates User interaction content, Indicates The number of interactions of the user's interactive content, express A collection of content tags. Content label The order of markings;

[0200] A content weight submodule, used to calculate the first content weight of the content hypergraph With the second content weight :

[0201] ;

[0202] ;

[0203] In the formula, Indicates user The total number of interactions, Indicates the number of users, Indicates user With content tags The number of interactions;

[0204] The iterative submodule is used to select a superpoint in the user hypergraph to iteratively transfer to the adjacent hyperedge, and calculate the first iteration vector after the iterative transfer , select a superpoint in the content hypergraph to iteratively transfer toward the adjacent hyperedge, and calculate the second iteration vector after iterative transfer :

[0205] ;

[0206] ;

[0207] ;

[0208] ; ;

[0209] ; ;

[0210] In the formula, , Represent the number of superpoints in the user hypergraph and content hypergraph respectively, is the iteration factor, , They represent the hyperedge vectors in the user hypergraph and content hypergraph respectively. denote the first incidence matrix and the second incidence matrix respectively, Respectively The elements in They represent the first Super point, A super edge, Respectively represent the first Super point, A super edge, , , , denote the first user diagonal matrix, the second user diagonal matrix, the first content diagonal matrix, and the second content diagonal matrix, respectively. Respectively , , , The elements in , Respectively represent The first iteration vector after iterations, , Respectively represent The second iteration vector after iterations;

[0211] A repeating submodule, used for repeating the iterative transfer process until an iterative condition is satisfied, and outputting a corresponding first iterative vector and a second iterative vector as a first final vector and a second final vector respectively;

[0212] The content scoring submodule is used to determine the user content score based on the first final vector and the second final vector.

[0213] The content scoring submodule includes:

[0214] A first vector unit is used to calculate the user vector based on the first final vector and the second final vector. With content vector :

[0215] ;

[0216] ;

[0217] In the formula, , Representation Matrix , Middle Line The elements of the column, , Representation Matrix , Middle Line The elements of the column, represents the user-label matrix, express The adjacency matrix of , Indicates the first final vector, the second final vector elements, Indicates the number of content tags;

[0218] The second vector unit is used to With content vector Calculate user score vector and content score vector :

[0219] ; ;

[0220] In the formula, , Set the user diagonal matrix for the first and second respectively, are the first and second setting content diagonal matrices respectively;

[0221] Content scoring unit, used to score based on user score vector and content score vector Determining User Content Ratings :

[0222] .

[0223] The second sorting module 2 comprises:

[0224] The evaluation value submodule is used to extract the evaluation information of the target user and the neighboring users from the content preference information, and determine the first target evaluation value according to the evaluation information , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value :

[0225] ; ;

[0226] ; ;

[0227] In the formula, represents the time control factor, is the maximum timestamp, , Respectively represent the target users With neighbor users User interaction content Evaluation, Indicate whether you like or dislike the review. Indicates the target user In the user interaction content Timestamps for likes and dislikes, Indicates neighbor users In the user interaction content Timestamps for likes and dislikes;

[0228] Similarity submodule, used to evaluate the value of the first target , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value Calculating user similarity :

[0229] ;

[0230] ;

[0231] In the formula, Represents user interaction content The comprehensive evaluation value of

[0232] The reliability submodule is used to classify the target user and neighbor users' preferences for the user interaction content into indicators to obtain several layers of indicator sets, determine the membership corresponding to the several layers of indicator sets, and calculate the first reliability value of the target user based on the membership The second most reliable value with neighbor users :

[0233] ;

[0234] ;

[0235] In the formula, Indicates the target user and neighbor user's interaction content The membership degree of , It represents the average membership degree of each layer indicator set corresponding to the target user and neighbor users. , Indicates the upper and lower limits of the user evaluation interval, , Indicates the target user and neighbor user's interaction content The actual evaluation value of

[0236] The time series scoring submodule is used to calculate the first reliable value of the target user. The second most reliable value with neighbor users Determine temporal content rating.

[0237] The temporal scoring submodule includes:

[0238] A reliability unit for determining a reliability value based on a first reliability value. The second reliable value Calculation time reliability :

[0239] ;

[0240] In the formula, , Represents the evaluation set of the target user and neighbor users;

[0241] Aggregate unit for time-based reliability Determine the set of neighbor users :

[0242] ;

[0243] In the formula, represents the initial neighbor user set, is the reliability threshold;

[0244] Temporal scoring unit, used to score based on neighbor user sets Determining Timed Content Ratings :

[0245] .

[0246] The third sorting module 3 comprises:

[0247] The keyword submodule is used to extract the user's search keyword from the content search information and calculate the similarity between the search keyword and the candidate keywords in the candidate keyword set. :

[0248] ;

[0249] In the formula, Indicates The search keyword corresponds to , Tag categories, Indicates The candidate keywords correspond to , Tag categories, , express , The common nodes of , express , The depth at which represents the distance calculation, , Respectively represent the number of label categories corresponding to the search keywords and candidate keywords;

[0250] A removal submodule is used to remove candidate keywords whose similarity is less than a similarity threshold from the candidate keyword set to obtain a removal keyword set;

[0251] Membership submodule, used to determine target users For the first membership degree of keywords in the keyword set and user interaction content For the second membership degree of keywords in the keyword set ;

[0252] Semantic submodule is used to identify The fuzzy set of the eliminated keywords is converted into several first semantic vectors, and the first semantic vectors are averaged to obtain the first average vector , based on the user interaction content The fuzzy set of the eliminated keywords is converted into several second semantic vectors, and the average of several second semantic vectors is taken to obtain the second average vector ;

[0253] Keyword scoring submodule, used to score keywords based on the first degree of membership , the second membership , the first mean vector , the second mean vector Calculate the first keyword score vector and the second keyword score vector :

[0254] ;

[0255] ;

[0256] Search scoring submodule, used to score vector based on the first keyword and the second keyword score vector Calculate content search score:

[0257] ;

[0258] In the formula, Represents cosine similarity.

[0259] The final sorting module 4 comprises:

[0260] A first weight submodule, configured to assign first weights to the to-be-recommended contents in the first sorted content set in a descending relationship according to the arrangement relationship of the first sorted content set;

[0261] A second weight submodule, configured to assign second weights to the to-be-recommended content in the second sorted content set in a descending relationship according to the arrangement relationship of the second sorted content set;

[0262] A third weight submodule, configured to assign a third weight to the content to be recommended in the third sorted content set in a descending relationship according to the arrangement relationship of the third sorted content set;

[0263] A comprehensive submodule, configured to calculate the sum of the first weight, the second weight, and the third weight of the same content to be recommended in the first sorted content set, the second sorted content set, and the third sorted content set, so as to obtain a comprehensive weight;

[0264] The final sorting submodule is used to sort the contents to be recommended in the content set to be recommended in descending order according to the size of the comprehensive weights to obtain a final sorted content set.

[0265] In some other embodiments of the present invention, the embodiments of the present invention provide the following technical solution: a computer, comprising a memory 102, a processor 101, and a computer program stored on the memory 102 and executable on the processor 101, wherein the processor 101 implements the content recommendation method based on the local life service platform as described above when executing the computer program.

[0266] Specifically, the processor 101 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiment of the present invention.

[0267] Among them, the memory 102 may include a large-capacity memory for data or instructions. By way of example and not limitation, the memory 102 may include a hard disk drive (HDD), a floppy disk drive, a solid state drive (SSD), a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 102 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 102 may be inside or outside the data processing device. In a specific embodiment, the memory 102 is a non-volatile memory. In a specific embodiment, the memory 102 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0268] The memory 102 may be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 101 .

[0269] The processor 101 implements the above-mentioned content recommendation method based on the local life service platform by reading and executing computer program instructions stored in the memory 102.

[0270] In some embodiments, the computer may further include a communication interface 103 and a bus 100. Figure 3 As shown, the processor 101, the memory 102, and the communication interface 103 are connected via a bus 100 and communicate with each other.

[0271] The communication interface 103 is used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present invention. The communication interface 103 can also implement data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage, and image / data processing workstations.

[0272] The bus 100 includes hardware, software or both, and couples the components of the computer device to each other. The bus 100 includes but is not limited to at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example and not limitation, bus 100 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, bus 100 may include one or more buses. Although embodiments of the present invention describe and illustrate a particular bus, the present invention contemplates any suitable bus or interconnect.

[0273] The computer can execute the content recommendation method based on the local life service platform of the present invention based on the content recommendation system based on the local life service platform, thereby realizing content recommendation based on the local life service platform.

[0274] In some further embodiments of the present invention, in combination with the above-mentioned content recommendation method based on the local life service platform, the embodiments of the present invention provide the following technical solutions: a storage medium having a computer program stored thereon, and the computer program implements the above-mentioned content recommendation method based on the local life service platform when executed by a processor.

[0275] Those skilled in the art will appreciate that the logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable instructions for implementing logical functions, and may be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For purposes of this specification, "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0276] More specific examples of readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0277] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or a combination thereof: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0278] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0279] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A content recommendation method based on a local life service platform, characterized in that: include: Obtaining content interaction information of a target user on a local life service platform, determining a user content score based on the content interaction information, and sorting the to-be-recommended content in the to-be-recommended content set in descending order based on the user content score to obtain a first sorted content set; determining content preference information based on the content interaction information, determining a time-series content score based on the content preference information, and sorting the to-be-recommended content in the to-be-recommended content set in descending order based on the time-series content score to obtain a second sorted content set; Obtaining content search information of the target user on the local life service platform, determining a content search score based on the content search information, and sorting the to-be-recommended content in the to-be-recommended content set in descending order based on the content search score to obtain a third sorted content set; The first sorted content set, the second sorted content set, and the third sorted content set are weightedly merged to obtain a final sorted content set, and the corresponding to-be-recommended content is recommended to the corresponding target user according to the sorting relationship in the final sorted content set; The step of determining the user content score based on the content interaction information comprises: Extracting the number of interactions and content labels from the content interaction information, constructing a user hypergraph with users as hyperedges and content labels as hyperpoints, and constructing a content hypergraph with user interaction content as hyperedges and content labels as superpoints; Calculate the first user weight of the user hypergraph With the second user weight : ； ； In the formula, Indicates the amount of user interaction content, Indicates User interaction content, Indicates The number of interactions of the user's interactive content, express A collection of content tags. Content label The order of markings; Calculate the first content weight of the content hypergraph With the second content weight : ； ； In the formula, Indicates user The total number of interactions, Indicates the number of users, Indicates user With content tags The number of interactions; In the user hypergraph, a hyperpoint is selected to iteratively transfer toward the adjacent hyperedge, and the first iteration vector after the iterative transfer is calculated. , select a superpoint in the content hypergraph to iteratively transfer toward the adjacent hyperedge, and calculate the second iteration vector after iterative transfer : ； ； ； ； ； ； ； In the formula, , Represent the number of superpoints in the user hypergraph and content hypergraph respectively, is the iteration factor, , They represent the hyperedge vectors in the user hypergraph and content hypergraph respectively. denote the first incidence matrix and the second incidence matrix respectively, Respectively The elements in They represent the first Super point, A super edge, Respectively represent the first superpoints, mth superedges, , , , denote the first user diagonal matrix, the second user diagonal matrix, the first content diagonal matrix, and the second content diagonal matrix, respectively. Respectively , , , The elements in , Respectively represent The first iteration vector after iterations, , Respectively represent The second iteration vector after iterations; Repeat the iterative transfer process until the iterative condition is met, and output the corresponding first iterative vector and second iterative vector as the first final vector and the second final vector respectively; Determine a user content score based on the first final vector and the second final vector; The step of determining the user content score based on the first final vector and the second final vector comprises: Calculate the user vector based on the first final vector and the second final vector With content vector : ； ； In the formula, , Representation Matrix , Middle Line The elements of the column, , Representation Matrix , Middle Line The elements of the column, represents the user-label matrix, express The adjacency matrix of , Indicates the first final vector, the second final vector elements, Indicates the number of content tags; Based on user vector With content vector Calculate user score vector and content score vector : ； ； In the formula, , Set the user diagonal matrix for the first and second respectively, are the first and second setting content diagonal matrices respectively; Based on user score vector and content score vector Determining User Content Ratings : 。 2. The content recommendation method based on the local life service platform according to claim 1, characterized in that: The step of determining the temporal content score based on the content preference information comprises: Extract evaluation information of the target user and neighbor users from the content preference information, and determine a first target evaluation value based on the evaluation information , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value : ； ； ； ； In the formula, represents the time control factor, is the maximum timestamp, , Respectively represent the target users With neighbor users User interaction content Evaluation, Indicates whether you like or dislike the review. Indicates the target user In the user interaction content Timestamps for likes and dislikes, Neighbor user In the user interaction content Timestamps for likes and dislikes; Based on the first target evaluation value , the second target evaluation value , first neighbor evaluation value , the second neighbor evaluation value Calculating user similarity : ； ； In the formula, Represents user interaction content The comprehensive evaluation value of The target user and neighbor users' preferences for the user interaction content are divided into index layers to obtain several layers of index sets, the membership degrees corresponding to the several layers of index sets are determined, and the first reliability value of the target user is calculated based on the membership degrees. The second most reliable value with neighbor users : ； ； In the formula, Indicates the target user and neighbor user's interaction content The membership degree of , It represents the average membership degree of each layer indicator set corresponding to the target user and neighbor users. , Indicates the upper and lower limits of the user evaluation interval, , Indicates the target user and neighbor user's interaction content The actual evaluation value of First reliable value based on target user The second most reliable value with neighbor users Determine temporal content rating.

3. The content recommendation method based on the local life service platform according to claim 2, characterized in that: The first reliability value based on the target user The second most reliable value with neighbor users The steps to determine the temporal content rating include: Based on the first reliable value The second reliable value Calculation time reliability : ； In the formula, , Represents the evaluation set of the target user and neighbor users; Time-based reliability Determine the set of neighbor users : ； In the formula, represents the initial neighbor user set, is the reliability threshold; Based on neighbor user set Determining Timed Content Ratings : 。 4. The content recommendation method based on the local life service platform according to claim 1, characterized in that: The step of determining a content search score based on the content search information comprises: Extract the user's search keyword from the content search information, and calculate the similarity between the search keyword and the candidate keywords in the candidate keyword set : ； In the formula, Indicates The search keyword corresponds to , Tag categories, Indicates The candidate keywords correspond to , Tag categories, , express , The common nodes of , express , The depth at which represents the distance calculation, , Respectively represent the number of label categories corresponding to the search keywords and candidate keywords; Eliminate candidate keywords whose similarity is less than a similarity threshold from the candidate keyword set to obtain an eliminated keyword set; Identify target users For the first membership degree of keywords in the keyword set and user interaction content For the second membership degree of keywords in the keyword set ; According to target users The fuzzy set of the eliminated keywords is converted into several first semantic vectors, and the first semantic vectors are averaged to obtain the first average vector , based on the user interaction content The fuzzy set of the eliminated keywords is converted into several second semantic vectors, and the average of several second semantic vectors is taken to obtain the second average vector ; Based on the first membership , the second membership , the first mean vector , the second mean vector Calculate the first keyword score vector and the second keyword score vector : ； ； Based on the first keyword score vector and the second keyword score vector Calculate content search score: ； In the formula, Represents cosine similarity.

5. The content recommendation method based on the local life service platform according to claim 1, characterized in that: The step of weightedly fusing the first sorted content set, the second sorted content set, and the third sorted content set to obtain a final sorted content set includes: assigning first weights to the to-be-recommended content in the first sorted content set in descending order according to the arrangement relationship of the first sorted content set; assigning second weights to the to-be-recommended content in the second sorted content set in descending order according to the arrangement relationship of the second sorted content set; assigning third weights to the to-be-recommended content in the third sorted content set in descending order according to the arrangement relationship of the third sorted content set; Calculating the sum of the first weight, the second weight, and the third weight of the same content to be recommended in the first ranked content set, the second ranked content set, and the third ranked content set to obtain a comprehensive weight; The contents to be recommended in the content set to be recommended are sorted in descending order according to the size of the comprehensive weights to obtain a final sorted content set.

6. A content recommendation system based on a local life service platform, the system adopts the content recommendation method based on a local life service platform as claimed in claim 1, characterized in that: The system comprises: A first sorting module is used to obtain content interaction information of a target user on a local life service platform, determine a user content score based on the content interaction information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the user content score to obtain a first sorted content set; a second sorting module, configured to determine content preference information based on the content interaction information, determine a time-series content score based on the content preference information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the time-series content score to obtain a second sorted content set; A third sorting module is used to obtain content search information of the target user on the local life service platform, determine a content search score based on the content search information, and sort the to-be-recommended content in the to-be-recommended content set in descending order based on the content search score to obtain a third sorted content set; The final sorting module is used to perform weighted fusion on the first sorted content set, the second sorted content set, and the third sorted content set to obtain a final sorted content set, and recommend the corresponding to-be-recommended content to the corresponding target user according to the sorting relationship in the final sorted content set.

7. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the content recommendation method based on the local life service platform as described in any one of claims 1 to 5.

8. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the content recommendation method based on the local life service platform as described in any one of claims 1 to 5.

Images