Information recommendation method and device

Abstract

The embodiment of the present application discloses an information recommendation method and device for improving the accuracy of information recommendation. The method includes: obtaining user characteristics of a target user and information characteristics of information; recommending information to the target user based on the similarity between the information characteristics and the user characteristics; wherein the information characteristics are obtained based on the correlation of information in information browsing records of multiple users; and the user characteristics are obtained based on the information characteristics of each information in the information browsing records of the corresponding user within a first preset time period.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to an information recommendation method and device. Background Art

[0002] Currently, Internet technology is developing rapidly and the amount of information on the Internet is huge. Therefore, how to enable users to quickly find the network information they want or that suits them and recommend suitable network information to users has become an urgent problem that technical personnel need to solve.

[0003] There are various ways to recommend information in the prior art, among which information can be recommended based on the relevance of tags, categories or keywords of the information and the browsing history of the user. However, this recommendation method is not accurate and detailed enough, has large errors, and has poor recommendation effect. Summary of the invention

[0004] In view of this, the embodiments of the present application provide an information recommendation method and device, which can solve the problem in the prior art that the recommendation results are not accurate and detailed enough and the recommendation effect is not good.

[0005] To solve the above problems, the technical solutions provided in the embodiments of the present application are as follows:

[0006] An information recommendation method, the method comprising:

[0007] Obtain user characteristics of target users and information characteristics of information;

[0008] Recommending information to the target user based on the similarity between the information feature and the user feature;

[0009] The information feature is obtained based on the correlation of information in the information browsing records of multiple users; and the user feature is obtained based on the information feature of each piece of information in the information browsing records of the corresponding user within a first preset time period.

[0010] In a possible implementation, the information feature includes a context feature of the corresponding information in a reference text; and the reference text is obtained based on information browsing records of the multiple users.

[0011] In a possible implementation, the information feature is obtained by using the following steps:

[0012] Acquiring information browsing records of multiple users within a second preset time period;

[0013] Based on the information browsing records of multiple users, an information relationship network is constructed; wherein the nodes in the information relationship network correspond to the information in the information browsing records one by one, and the line between two nodes in the information relationship network represents that the two information corresponding to the two nodes appear in the information browsing records of the same user at the same time;

[0014] According to preset walking parameters, performing at least one random walk in the information relationship network to obtain various walking paths;

[0015] Constructing the reference text based on each obtained wandering path;

[0016] Obtaining context features of the information corresponding to the nodes in the information relationship network in the reference text;

[0017] Based on the context feature, the information feature of the corresponding information is obtained.

[0018] In a possible implementation, a line between two nodes in the information relationship network represents that the information corresponding to the two nodes simultaneously appears in the information browsing record of the same user, and the information corresponding to the two nodes is operated by the same user in sequence;

[0019] Alternatively, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing history of the same user, and the information corresponding to the two nodes is operated in sequence by the same user; the connection between two nodes in the information relationship network has a direction pointing to one of the nodes, and the direction of the connection is determined by the order in which the same user operates on the information corresponding to the two nodes connected by the connection.

[0020] In a possible implementation, a line between two nodes in the information relationship network represents that the information corresponding to the two nodes simultaneously appears in the information browsing record of the same user, and the time interval between the operation times of the same user on the information corresponding to the two nodes is less than a first preset threshold;

[0021] And / or, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing record of the same user, and the amount of information browsed by the same user between browsing the information corresponding to the two nodes is less than a second preset threshold.

[0022] In a possible implementation, constructing the reference text based on each obtained walking path specifically includes:

[0023] Arrange the nodes in each of the walking paths in order to obtain a first text;

[0024] The nodes in the first text are replaced with an information parameter of corresponding information to obtain the reference text.

[0025] In a possible implementation, when the context feature is a feature vector and there are multiple features, obtaining the information feature of the corresponding information based on the context feature specifically includes:

[0026] Concatenating the context features of each information parameter in the reference text to obtain the information features of the corresponding information;

[0027] Alternatively, the context feature of each information parameter in the reference text is multiplied by the corresponding weight and then concatenated to obtain the information feature of the corresponding information.

[0028] In a possible implementation manner, when the context feature is a feature vector and there are multiple context features, the dimensions of the context features corresponding to different information parameters are not all the same.

[0029] In a possible implementation, for the cold information that is not included in the information browsing records of the multiple users, the information features of the cold information are obtained by using the following steps:

[0030] Determine reference information in the information browsing records of the multiple users that has the same information parameters as the cold information;

[0031] The information feature of the cold information is obtained according to the information feature of the reference information.

[0032] In a possible implementation, the information parameter is an identity identifier, a category, a source, a keyword, or a personalized tag.

[0033] In a possible implementation, when the information feature is a feature vector, the user feature is obtained by using the following steps:

[0034] taking the average value of the information feature of each information in the information browsing record of the corresponding user within the first preset time period as the user feature of the corresponding user;

[0035] Alternatively, the average value of the information feature of each information in the information browsing record of the corresponding user within the first preset time period multiplied by the corresponding coefficient is used as the user feature of the corresponding user.

[0036] In a possible implementation manner, when both the information feature and the user feature are feature vectors, recommending information to the target user according to the similarity between the information feature and the user feature specifically includes:

[0037] Obtaining a recommendation parameter according to the distance between the user feature of the target user and the information feature of the information to be recommended; the recommendation parameter is negatively correlated with the distance;

[0038] Determining whether the recommended parameter is greater than a third preset threshold;

[0039] When the recommendation parameter is greater than the third preset threshold, the information to be recommended is recommended to the target user.

[0040] In a possible implementation manner, when both the information feature and the user feature are feature vectors, recommending information to the target user according to the similarity between the information feature and the user feature specifically includes:

[0041] According to the distance between the user feature of the target user and the information feature of each piece of information to be recommended in the set of information to be recommended, a recommendation parameter of each piece of information to be recommended is obtained; the recommendation parameter is negatively correlated with the distance;

[0042] The recommendation parameters of each of the information to be recommended are sorted, and the information to be recommended corresponding to the largest first N recommendation parameters is recommended to the target user; N is a positive integer.

[0043] An information recommendation device, comprising:

[0044] A feature acquisition module is used to acquire user features of target users and information features of information;

[0045] An information recommendation module, used to recommend information to the target user based on the similarity between the information feature and the user feature;

[0046] The information feature is obtained based on the correlation of information in the information browsing records of multiple users; and the user feature is obtained based on the information feature of each piece of information in the information browsing records of the corresponding user within a first preset time period.

[0047] A device for information recommendation includes a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors. The one or more programs include instructions for performing the following operations:

[0048] Obtain user characteristics of target users and information characteristics of information;

[0049] Recommending information to the target user based on the similarity between the information feature and the user feature;

[0050] The information feature is obtained based on the correlation of information in the information browsing records of multiple users; and the user feature is obtained based on the information feature of each piece of information in the information browsing records of the corresponding user within a first preset time period.

[0051] A computer-readable medium stores instructions which, when executed by one or more processors, enable a device to perform the information recommendation method as described above.

[0052] Compared with the prior art, this application has at least the following advantages:

[0053] In an embodiment of the present application, the information characteristics of each information can be determined based on the information browsing records of multiple users, and the user characteristics of the target user can be obtained based on the information characteristics of each information in the information browsing records of the target user within the first preset time period. Then, according to the similarity between the information characteristics of the information and the user characteristics of the target user, the information can be accurately recommended to the target user. Because the information characteristics reflect the relevance of the information with other information in the information browsing records of multiple users, and the user characteristics of the target user comprehensively reflect the information characteristics of the target user's browsing information, representing the information browsing tendency and trend of the target user, when the similarity between the information characteristics and the user characteristics is high, it means that the information corresponding to the information characteristics is consistent with the user's information browsing tendency and trend, and recommending the information to the target user can improve the accuracy of information recommendation. In addition, the information recommendation method and device provided in the embodiment of the present application are not limited by information categories or tags, and the recommended information is diverse, which improves the effect of information recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, 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 recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0055] Figure 1 A schematic diagram of a system framework involved in an application scenario in an embodiment of the present application;

[0056] Figure 2 A flowchart of an information recommendation method provided in an embodiment of the present application;

[0057] Figure 3 A flowchart of an information recommendation method provided for a specific embodiment of the present application;

[0058] Figure 4 A flowchart of another information recommendation method provided in a specific embodiment of the present application;

[0059] Figure 5 A flowchart of a method for acquiring information features provided in an embodiment of the present application;

[0060] Figure 6a-6e A schematic diagram of the structure of the information relationship network provided for a specific embodiment of the present application;

[0061] Figure 7 A flowchart of another information feature acquisition method provided in an embodiment of the present application;

[0062] Figure 8 A schematic diagram of the structure of an information recommendation device provided in an embodiment of the present application;

[0063] Fig. 9 A schematic diagram of the structure of a device for information recommendation provided in an embodiment of the present application;

[0064] Fig.10 A schematic diagram of the structure of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0065] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0066] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0067] In traditional information recommendation methods, recommendations are generally made based on information labels, categories or keywords. When the labels, categories or keywords of the information to be recommended are the same as those of the information that the target user has browsed, the information to be recommended is recommended to the target user. However, the inventors of this application have found in their research that due to the coarse granularity of the classification of information labels, categories or keywords, the existing information recommendation methods are not accurate or detailed enough, and there is no diversity and extensibility in information recommendation, and the effect of information recommendation is not good.

[0068] Taking the recommended information as news as an example, when the target user has browsed the listing news about Company A (that is, the information that the target user has browsed), in the traditional information recommendation method, because the labels of the listing news of Company A are set as economy and Company A, the label setting is not precise enough. When recommending information, all news labeled as economy and Company A will be recommended to the target user. However, the target user is actually concerned about the listing news related to Company A, resulting in the information recommendation being not precise and detailed enough, and it is also impossible to recommend listing news about other companies or other news about Company A or economic news of other companies to the target user. There is a lack of diversity and expansibility in information recommendation, and it is impossible to meet the information browsing needs of the target user, and the effect of information recommendation is poor.

[0069] In order to solve this problem, the embodiment of the present application provides an information recommendation method and device, which determines the information characteristics of information based on the information browsing records of multiple users, and the information characteristics can reflect the correlation between the corresponding information and other information in the user's browsing record, and based on the information characteristics of each information in the target user's information browsing record within the first preset time period, the user characteristics of the target user can be determined, which can reflect the information browsing tendency and trend of the target user. Therefore, based on the similarity between the information characteristics of the information and the user characteristics of the target user, it can be reflected whether the information is related to the information that the target user has browsed, and whether it conforms to the user's information browsing tendency and trend. Then, according to the similarity between the information characteristics of the information and the user characteristics of the target user, the information can be accurately recommended to the target user. In addition, the information recommendation method and device provided by the embodiment of the present application are not limited by information categories or tags, and the recommended information is diverse, which improves the effect of information recommendation.

[0070] For example, one of the scenarios of the embodiments of the present application may be applied to Figure 1 In the scenario shown. The scenario may include a server 101 and a terminal 102. A user may use the terminal 102 to browse information, and the server 101 obtains information features of the information and user features of the target user based on the information browsing records of multiple users and the information browsing records of the target user, and recommends information to the target user based on the information features of the information and the user features of the target user.

[0071] In the above application scenario, although the actions of the implementation mode of the present application are described as being executed by the server 101, these actions may also be executed by the client 102, or may be partially executed by the client 102 and partially executed by the server 101. The present application is not limited in terms of the execution subject, as long as the actions disclosed in the implementation mode of the present application are executed. Those skilled in the art can understand that Figure 1 The framework diagram shown is only an example in which the embodiments of the present application can be implemented. The scope of application of the embodiments of the present application is not limited by any aspect of the framework.

[0072] It should be noted that the terminal 102 in the embodiment of the present application can be any user device that is currently available, under development or will be developed in the future and can interact with the server 101 through any form of wired and / or wireless connection (e.g., Wi-Fi, LAN, cellular, coaxial cable, etc.), including but not limited to: smart phones, non-smart phones, tablet computers, laptop personal computers, desktop personal computers, small computers, medium-sized computers, large computers, etc. that are currently available, under development or will be developed in the future. It should also be noted that the server 101 in the embodiment of the present application can be an example of a device that is currently available, under development or will be developed in the future and can provide information recommendation services to users. The implementation methods of the present application are not limited in this regard.

[0073] It should be noted that the information recommendation method and device provided in the embodiments of the present application are applicable to various types of information recommendations, such as personalized recommendations. The recommended information includes but is not limited to: news, commodities, audio and video, websites, entertainment activities, books, articles and applications.

[0074] The information recommendation method provided by the embodiment of the present application will be described in detail below with reference to the accompanying drawings.

[0075] See also Figure 2 , which is a flow chart of an information recommendation method provided in an embodiment of the present application.

[0076] The information recommendation method provided in the embodiment of the present application includes:

[0077] S201: Acquire user characteristics of a target user and information characteristics of the information.

[0078] In the embodiment of the present application, the information feature is obtained based on the correlation of information in the information browsing records of multiple users, and the user feature is obtained based on the information feature of each piece of information in the information browsing record of the corresponding user within the first preset time period. It should be noted that the information browsing record includes but is not limited to any one or more of user clicks, browsing, collections, and likes. As long as the user performs an operation on the information, the information can be included in the information browsing record. In the following, operation refers to the behavior performed by the user on each piece of information in his or her information browsing record. The operation includes but is not limited to clicks, browsing, collections, likes, purchases, etc., which are not listed here one by one.

[0079] The information features of the target information can reflect the correlation between the target information and other information that the user has operated when the target information is operated by the user. The user features are obtained based on the information features of each information in the information browsing record of the corresponding user within the first preset time period, and can reflect the user's information operation tendency and / or trend. Therefore, the user features of the target user and the information features of the information can determine the correlation between the information and the target user's information operation tendency and / or trend, determine whether to recommend the information to the target user, and achieve accurate recommendation of information.

[0080] In some possible implementations of the embodiments of the present application, the information browsing records of multiple users can be regarded as a text (i.e., a reference text), because the context of the target information in the text is other information in the same information browsing record as the target information, and the context feature can be used to reverse the correlation between the target information and other information that the user has operated when operated by the user, so the information feature of the information may include the text feature of the information in the text. That is, the information feature includes the context feature of the corresponding information in the reference text; the reference text is obtained based on the information browsing records of multiple users. In practical applications, the context feature of the information in the reference text can be obtained by any method, such as a word vector algorithm, etc. The context feature can be represented by a feature vector, which is not limited here. The following will explain in detail how to obtain the information feature of the information in conjunction with a specific example, which will not be repeated here.

[0081] In some possible implementations of the embodiments of the present application, when the information feature is a feature vector, the user feature is obtained by using the following steps:

[0082] The average value of the information features of each piece of information in the information browsing record of the corresponding user within the first preset time period is used as the user feature of the corresponding user; or, the average value of the information features of each piece of information in the information browsing record of the corresponding user within the first preset time period multiplied by the corresponding coefficient is used as the user feature of the corresponding user.

[0083] In an embodiment of the present application, the information browsing record of the corresponding user may include the information browsed by the user in the most recent period of time (i.e., the first preset time period), or may only include the few pieces of information browsed by the user most recently, which is not limited here. It should be noted that the first preset time period can be set according to the specific application scenario and actual needs. Because the user's information operation interest has a certain timeliness, the first preset time period should generally not be set too long or too far away from the current time interval. As an example, assuming that the current time is the kth day, the first preset time period can be the k-1th day to the knth day, where n is a positive integer, such as n=15.

[0084] In practical applications, the coefficients corresponding to the information features of each piece of information can be set according to specific needs to determine the tendency of information recommendation, which will not be listed here. It can be understood that when the information features are represented by feature vectors, the user features are also feature vectors.

[0085] S202: Recommend information to the target user based on the similarity between the information feature and the user feature of the target user.

[0086] It is understandable that, since the information characteristics of information can reflect the correlation between the target information and other information that the user has browsed when operated by the user, and the user characteristics of the target user can reflect the information operation tendency and / or trend of the target user, the more similar the information characteristics of information and the user characteristics of the target user are, the more the information conforms to the information operation tendency and / or trend of the target user. Therefore, according to the similarity between the information characteristics of information and the user characteristics of the target user, it can be determined whether the information conforms to the information operation tendency and / or trend of the target user, and recommending information to the target user based on the similarity can improve the accuracy and precision of the information recommendation.

[0087] Moreover, when recommending information, parameters such as labels and categories are not used to restrict the recommended information. Instead, recommendations are made based on the relevance of the information to other information when it is operated and the target user's information operation inclinations and trends. On the one hand, the recommended information is diverse, and on the other hand, it can also adapt to the development of the target user's information operation interests to recommend information, thereby improving the effect of information recommendation.

[0088] In practical applications, any feature similarity calculation method can be used to obtain the similarity between the information feature and the user feature of the target user, such as using the distance between feature vectors to represent the feature similarity, etc., which is not limited here.

[0089] In practical applications, when both the information feature and the user feature are feature vectors, there are at least two possible implementations of step S202, which are described below one by one:

[0090] In the first possible implementation, Figure 3 As shown, step S202 may specifically include:

[0091] S301: Obtain recommendation parameters according to the distance between the user characteristics of the target user and the information characteristics of the information to be recommended.

[0092] In an embodiment of the present application, the recommendation parameter is negatively correlated with the distance between the user characteristics of the target user and the information characteristics of the information to be recommended, and the distance between the recommendation parameter and the user characteristics of the target user and the information characteristics of the information to be recommended can be represented by cosine similarity or Euclidean distance between feature vectors. It can be understood that when the cosine similarity between the user characteristics of the target user and the information characteristics of the information to be recommended is closer to 1, the distance between the two is smaller, and the recommendation parameter is larger; when the Euclidean distance between the user characteristics of the target user and the information characteristics of the information to be recommended is smaller, the recommendation parameter is larger. The specific form of the recommendation parameter can be set according to actual needs, and will not be listed here one by one.

[0093] S302: Determine whether the recommended parameter is greater than a third preset threshold; when the recommended parameter is greater than the third preset threshold, execute step S303.

[0094] S303: Recommend the information to be recommended to the target user.

[0095] It is understandable that in the first possible implementation, information whose recommendation parameter is greater than the third preset threshold is recommended to the target user. In practical applications, the third preset threshold and the specific recommendation method can be set according to actual needs and are not limited here.

[0096] In the second possible implementation, Figure 4 As shown, step S202 may specifically include:

[0097] S401: Obtain recommendation parameters of each piece of information to be recommended according to the distance between the user feature of the target user and the information feature of each piece of information to be recommended in the set of information to be recommended.

[0098] Similar to the first possible implementation, in the second possible implementation, the recommendation parameter is negatively correlated with the distance between the user characteristics of the target user and the information characteristics of the information to be recommended, and the distance between the recommendation parameter and the user characteristics of the target user and the information characteristics of the information to be recommended can be represented by cosine similarity or Euclidean distance between feature vectors. It can be understood that when the cosine similarity between the user characteristics of the target user and the information characteristics of the information to be recommended is closer to 1, the distance between the two is smaller, and the recommendation parameter is larger; when the Euclidean distance between the user characteristics of the target user and the information characteristics of the information to be recommended is smaller, the recommendation parameter is larger. The specific form of the recommendation parameter can be set according to actual needs, and will not be listed here one by one.

[0099] S402: Sort the recommendation parameters of each piece of information to be recommended, and recommend the information to be recommended corresponding to the largest first N recommendation parameters to the target user.

[0100] It is understandable that N is a positive integer. In the second possible implementation, the first N information with the largest recommendation parameters are recommended to the target user. In practical applications, the value of N and the specific recommendation method can be set according to actual needs, which is not limited here.

[0101] In an embodiment of the present application, the information characteristics of each information can be determined based on the information browsing records of multiple users, and the user characteristics of the target user can be obtained based on the information characteristics of each information in the information browsing records of the target user within the first preset time period. Then, according to the similarity between the information characteristics of the information and the user characteristics of the target user, the information can be accurately recommended to the target user. Because the information characteristics reflect the relevance of the information with other information in the information browsing records of multiple users, and the user characteristics of the target user comprehensively reflect the information characteristics of the target user's browsing information, representing the information browsing tendency and trend of the target user, when the similarity between the information characteristics and the user characteristics is high, it means that the information corresponding to the information characteristics is consistent with the information browsing tendency and trend of the target user, and recommending the information to the target user can improve the accuracy of information recommendation. In addition, the information recommendation method and device provided in the embodiment of the present application are not limited by information categories or tags, and the recommended information is diverse, which improves the effect of information recommendation.

[0102] The following describes in detail how to obtain the information characteristics of the information with reference to a specific example.

[0103] See also Figure 5 , which is a flow chart of an information feature acquisition method provided in a specific embodiment of the present application.

[0104] In some possible implementations of the embodiments of the present application, the information features can be obtained by using the following steps:

[0105] S501: Acquire information browsing records of multiple users within a second preset time period.

[0106] In the embodiment of the present application, the second preset time period may be equal to or different from the first preset time period, and may be set according to actual needs, which is not limited here. It is understandable that the information browsing records of multiple users in the second preset time period are used as references for information recommendation in the embodiment of the present application. In order to ensure the effect of information recommendation, it should be ensured that enough information browsing record data is obtained, such as a sufficient number of users and / or a sufficient amount of data in the information browsing record, which will not be listed one by one.

[0107] S502: Building an information relationship network based on the information browsing records of multiple users.

[0108] In the embodiment of the present application, the nodes in the information relationship network correspond one-to-one to the information in the information browsing records of multiple users, and the line between two nodes in the information relationship network represents that the corresponding two information appear in the information browsing records of the same user at the same time. That is, each node in the information relationship network represents one information in the information browsing records of multiple users obtained; when two information appear in the information browsing records of the same user, there is a line between the nodes corresponding to the two information; conversely, when the two information do not appear in the information browsing records of the same user, there is no line between the nodes corresponding to the two information. It can be understood that the lines in the information relationship network represent the correlation between the information browsed by the same user.

[0109] For example, the information browsing record of user U1 includes information A, B and C, the information browsing record of user U2 includes information C, B, F and D, and the information browsing record of user U3 includes information B and E, as shown in Table 1 below:

[0110] Table 1 Information browsing records of multiple users within the first preset time period

[0111]

[0112] Based on this, an information relationship network including nodes A, B, C, D, E and F can be constructed, and nodes A, B, C, D, E and F correspond to information A, B, C, D, E and F respectively.

[0113] Then, in some possible designs, because information A, B and C all appear in the information browsing record of user U1, there are connections between the corresponding nodes A, B and C in the information relationship network; because information C, B, F and D all appear in the information browsing record of user U2, there are connections between the corresponding nodes C, B, F and D in the information relationship network; because information B and E all appear in the information browsing record of user U3, there is a connection between the corresponding nodes B and E in the information relationship network. The constructed information relationship network can be as follows Figure 6a shown.

[0114] In some possible implementations of the embodiments of the present application, the connection method between nodes in the information relationship network can also be adaptively adjusted according to the actual application scenario and specific circumstances. In practical applications, there are at least the following two possible implementations:

[0115] In a first possible implementation, the development trend of users' information browsing, that is, the development of users' information operation interests, can be comprehensively considered, and the connections between nodes in the constructed information relationship network can be determined according to the order of users' information operations in the information browsing records. The determined connections can be directional or non-directional.

[0116] In one example, a line between two nodes in an information relationship network represents that the information corresponding to the two nodes simultaneously appears in the information browsing history of the same user, and the information corresponding to the two nodes is operated by the same user in sequence. That is, when two pieces of information are operated by the same user in sequence, the nodes corresponding to the two pieces of information are connected.

[0117] Continuing with the above example, assuming that the order of the information in the information browsing record in the above example is the order in which the user operates on the information, then in the constructed information relationship network, there are undirected connections between nodes A and B, nodes B and C, nodes B and E, nodes B and F, and nodes F and D, such as Figure 6b shown.

[0118] In another example, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears in the information browsing history of the same user at the same time, and the information corresponding to the two nodes is operated by the same user in sequence; the connection between two nodes in the information relationship network has a direction pointing to one of the nodes, and the direction of the connection is determined by the order in which the same user operates on the information corresponding to the two nodes connected by the connection. That is, when two pieces of information are operated by the same user in sequence, the nodes corresponding to the two pieces of information are connected, and the direction of the connection is from the node corresponding to the information operated first by the same user to the node corresponding to the information operated later by the user.

[0119] Continuing with the above example, assuming that the order of the information in the information browsing record in the above example is the order in which the user operates on the information, then in the constructed information relationship network, there are connections from node A to node B, from node B to node C, from node C to node B, from node B to node F, from node F to node D, and from node B to node E, such as Figure 6c shown.

[0120] In a second possible implementation, the timeliness of the user's information operation can also be considered. When determining the connection between nodes in the constructed information relationship network, for two information in the information browsing record of the same user, if the time interval between the user's browsing time of the two information or the number of information in the interval exceeds a certain threshold, then the nodes corresponding to the two information will not be connected; if the time interval between the user's browsing time of the two information or the number of information in the interval does not exceed a certain threshold, then there will be a connection between the nodes corresponding to the two information.

[0121] In one example, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears in the information browsing history of the same user at the same time, and the time interval between the operation times of the same user on the information corresponding to the two nodes is less than the first preset threshold. That is, when the time interval between the operation times of the same user on the two information is less than the first preset threshold, there is a connection between the corresponding nodes of the two information in the information relationship network. The connection can be without direction, or it can have a direction determined by the operation sequence of the same user on the information.

[0122] Continuing with the above example, assuming that the order of the information in the information browsing record in the above example is the order in which the user operates on the information, and only the time interval between the operation time of user U2 on information C and information D exceeds the first preset threshold, then in the constructed information relationship network, there are connections between nodes A, B and C, between nodes C, B and F, between nodes B, F and D, and between nodes B and E, such as Figure 6d shown.

[0123] In practical applications, the first preset threshold can be set according to specific application scenarios and actual needs, and is not limited here.

[0124] In some possible designs, the connection in the information relationship network may exist only between nodes corresponding to two information operated sequentially by the same user, or may exist between nodes corresponding to two information operated by the same user for a time less than a first preset threshold. For details, please refer to the description in the first possible implementation method above and will not be repeated here.

[0125] In another example, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears in the information browsing record of the same user at the same time, and the number of information browsed by the same user between browsing the information corresponding to the two nodes is less than the second preset threshold. That is, when the same user operates on two pieces of information, the same user also operates on less than the second preset threshold, there is a connection between the corresponding nodes of the two pieces of information in the information relationship network.

[0126] Continuing with the above example, assuming that the order of the information in the information browsing record in the above example is the order in which the user operates on the information, and the second preset threshold is 1, then in the constructed information relationship network, there is no connection between the nodes corresponding to information C and information D, there are connections between nodes A, B and C, there are connections between nodes C, B and F, there are connections between nodes B, F and D, and there is a connection between nodes B and E. Figure 6d shown.

[0127] In practical applications, the second preset threshold can be set according to specific application scenarios and actual needs, and is not limited here.

[0128] In some possible designs, when constructing a relationship network, it is not only possible to consider that the number of information operated between the operations of the same user on two pieces of information does not exceed the second preset threshold, but it is also possible to simultaneously consider that the time interval between the operations of the same user on the two pieces of information does not exceed the first preset threshold. For specific implementations, please refer to the relevant descriptions above. Similarly, the connection in the information relationship network may only exist between the nodes corresponding to the two pieces of information operated sequentially by the same user, or between the nodes corresponding to the two pieces of information operated by the same user for a time period less than the first preset threshold. For details, please refer to the description in the first possible implementation method above, and no further elaboration is required.

[0129] It should also be noted that the above content illustrates various ways of constructing information relationship networks. This is because the lines between nodes in the information relationship network represent the correlation between the corresponding information, which affects the expression of information features. Therefore, in practical applications, in order to improve the accuracy and effectiveness of information recommendation, it is possible to determine whether the correlation between information has a logical relationship in browsing time and order according to the specific application scenario, and based on this, determine whether to connect the corresponding nodes when constructing the information relationship network.

[0130] For example, when the recommended information is news or goods, the user's browsing interest in news or goods has a certain timeliness and sequence. When constructing the information relationship network, the timeliness and sequence of the user's browsing interest in news or goods can be considered to determine whether there is a connection between the nodes in the information relationship network, such as using the first implementation method or the second implementation method mentioned above to construct the information relationship network. For another example, when the recommended information is music, the user's interest is relatively fixed. When constructing the information relationship network, two pieces of information that appear in the same user's information browsing record at the same time are directly connected at the corresponding nodes in the information relationship network, ignoring the timeliness and sequence of the user's interest.

[0131] As can be seen from the above examples, two of the information may appear in the information browsing records of multiple users. For example, in the example of Table 1, information B and information C are included in the information browsing records of user U1 and user U2. Then, in some possible implementations, a weight can be set for the connection between the corresponding two nodes based on the number of times the two information appear in the information browsing records of multiple users, so as to better reflect the correlation between the information and improve the accuracy of information recommendation. It can be understood that the higher the weight of the connection, the greater the correlation between the two information corresponding to the connection. Figure 6e Taking the example shown in Table 1 as an example, an information relationship network including connection weights is illustrated, and the numbers on the connection lines represent their corresponding weights.

[0132] S503: Perform at least one random walk in the information relationship network according to preset walk parameters to obtain various walk paths.

[0133] In an embodiment of the present application, a random walk is to start from any node in the information relationship network and perform a random walk traversal along the lines between the nodes. The walk parameters can set the longest number of steps of the walk (corresponding to the number of nodes passed), the number of times each node is used as the starting point, the probability of walking to each node, etc. For example, the longest number of walks can be set to 4, and the number of times each node is used as the starting point can be set to 2. For another example, the probability of walking to the node of the previous walk again, the probability of walking to other nodes connected to the node of the previous walk, and the probability of walking to other nodes can be set to set whether the random walk strategy is biased towards depth-first search or breadth-first search, thereby controlling the information recommendation strategy. It should be noted that when a weight is set on the line of the information relationship network, the weight will also affect the probability of the line being walked.

[0134] by Figure 6e Taking the information relationship network shown in the figure, the longest number of walking steps is set to 4, and the number of times each node is used as a starting point is set to 2 as an example, the following Table 2 shows examples of various walking paths obtained by multiple random walks.

[0135] Table 2 Multiple walking paths

[0136]

[0137] S504: Construct reference texts based on the obtained wandering paths.

[0138] In an embodiment of the present application, after arranging the nodes in each obtained wandering path in order, each node can be replaced with information parameters of the corresponding information (such as identification, source, category, etc.) to obtain a reference text to indicate the user's operation on the information.

[0139] In some possible designs, step S504 may specifically include:

[0140] The nodes in each walking path are arranged in order to obtain a first text; the nodes in the first text are replaced by an information parameter of corresponding information to obtain a reference text.

[0141] In the embodiment of the present application, the information parameter may specifically be an identity identifier, category, source, keyword (such as source website, author) or personalized tag, etc. In practical applications, the nodes in the first text may be replaced with the information parameters of the corresponding information to obtain multiple reference texts, and the following operations may be performed based on each reference text to obtain multiple context features of the information, and then each context feature may be combined to obtain the information feature of the corresponding information.

[0142] S505: Obtain context features of the information corresponding to the nodes in the information relationship network in the reference text.

[0143] In an embodiment of the present application, any context feature acquisition method can be used to obtain the context features of the information in the reference text, such as the skip-gram model of the word2vec framework, etc., which will not be repeated here. The obtained context features can be represented by feature vectors, and the dimension of the feature vector can be set according to actual needs. In some possible designs, when multiple reference texts are obtained based on multiple information parameters, a context feature corresponding to the information parameter is obtained for each reference text. In an example, different dimensions can be set for the context features corresponding to each information parameter according to actual needs, and the dimensions of any two context features obtained can be the same or different. For example, when the context feature is a feature vector and there are multiple context features, the dimension of the feature vector can be set based on the contribution of the information parameters corresponding to each context feature in information recommendation. The higher the contribution, the higher the dimension of the corresponding feature vector.

[0144] S506: Based on the context feature, obtain information features of the corresponding information.

[0145] In the embodiment of the present application, the context feature can be directly used as the information feature of the corresponding information, or the context feature can be multiplied by a certain coefficient or calculated based on a preset formula to obtain the information feature of the corresponding information, which is not limited here.

[0146] In some possible implementations of the embodiments of the present application, when the context feature is a feature vector and there are multiple context features, step S506 may specifically include:

[0147] The context features of each information parameter in the reference text are concatenated to obtain the information features of the corresponding information; or, the context features of each information parameter in the reference text are multiplied by the corresponding weights and then concatenated to obtain the information features of the corresponding information.

[0148] It can be understood that, assuming that three context features are obtained, namely (a, b, c), (d, e, f, g) and (h,i, j), the information features obtained by splicing can be (a, b, c, d, e, f, g, h, i, j). When the weights corresponding to each context feature are α, β and γ respectively, the information features obtained by splicing can be (αa, αb, αc, βd, βe, βf, βg, γh, γi, γj). In the embodiment of the present application, the weight can be used to set the tendency represented by the feature vector, and the weight corresponding to each context feature can be set according to actual needs, and they are not listed one by one here.

[0149] The above content describes in detail how to obtain the information features of each information in the information browsing records of multiple users. In practical applications, as the information increases, although the information features of each information can be updated in a timely manner according to the browsing information of the user, there may still be information that is not included in the information browsing records of the multiple users (i.e., cold information). In order to obtain the information features of the cold information so as to recommend the cold information to the user, in some possible implementations of the embodiments of the present application, the information features of the cold information can be obtained based on the information features of other information similar to the cold information.

[0150] Specifically, Figure 7 As shown, for cold information that is not included in the information browsing records of multiple users, the information features of the cold information are obtained by using the following steps:

[0151] S701: Determine reference information having the same information parameters as the cold information in the information browsing records of multiple users.

[0152] In the embodiment of the present application, the reference information and the cold information have at least one common information parameter, such as the same category, the same source, etc. It is understandable that the information parameter can be an identity identifier, category, source, keyword or personalized tag, etc., which is not limited here.

[0153] S702: Obtain information features of cold information according to information features of reference information.

[0154] In the embodiment of the present application, the information features of the cold information are identified by the information features of the reference information that are similar to the cold information. The method for obtaining the information features of the reference information can specifically adopt any of the above descriptions, which will not be repeated here. As an example, when there is one reference information, the information features of the reference information can be used as the information features of the cold information; when there are multiple reference information and the information features are feature vectors, the average value of the information features of each reference information can be used as the information features of the cold information, which is not limited here.

[0155] Based on the information recommendation method provided in the above embodiment, the embodiment of the present application also provides an information recommendation device.

[0156] See also Figure 8 , which is a structural diagram of an information recommendation device provided in an embodiment of the present application.

[0157] The information recommendation device provided in the embodiment of the present application includes:

[0158] The feature acquisition module 801 is used to acquire user features of the target user and information features of the information;

[0159] The information recommendation module 802 is used to recommend information to target users based on the similarity between information features and user features;

[0160] The information feature is obtained based on the correlation of information in the information browsing records of multiple users; the user feature is obtained based on the information feature of each information in the information browsing record of the corresponding user within the first preset time period.

[0161] In some possible implementations of the embodiments of the present application, the information features may specifically include context features of the corresponding information in the reference text; the reference text may be obtained based on the information browsing records of multiple users.

[0162] In some possible implementations of the embodiments of the present application, the feature acquisition module 801 may specifically include:

[0163] A record acquisition submodule, used to acquire information browsing records of multiple users within a second preset time period;

[0164] The network construction submodule is used to construct an information relationship network based on the information browsing records of multiple users; wherein the nodes in the information relationship network correspond to the information in the information browsing records one by one, and the connection between two nodes in the information relationship network represents that the two information corresponding to the two nodes appear in the information browsing records of the same user at the same time;

[0165] A random walk submodule is used to perform at least one random walk in the information relationship network according to preset walk parameters to obtain various walk paths;

[0166] A text construction submodule is used to construct reference texts based on each obtained wandering path;

[0167] The first acquisition submodule is used to obtain the contextual features of the information corresponding to the nodes in the information relationship network in the reference text;

[0168] The second acquisition submodule is used to obtain information features of corresponding information based on the context features.

[0169] In some possible implementations of the embodiments of the present application, a line between two nodes in the information relationship network represents that the information corresponding to the two nodes simultaneously appears in the information browsing record of the same user, and the information corresponding to the two nodes is operated by the same user in sequence;

[0170] Alternatively, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing history of the same user, and the information corresponding to the two nodes is operated in sequence by the same user; the connection between two nodes in the information relationship network has a direction pointing to one of the nodes, and the direction of the connection is determined by the order in which the same user operates on the information corresponding to the two nodes connected by the connection.

[0171] In some possible implementations of the embodiments of the present application, a line between two nodes in the information relationship network represents that the information corresponding to the two nodes simultaneously appears in the information browsing record of the same user, and the time interval between the operation times of the same user on the information corresponding to the two nodes is less than a first preset threshold;

[0172] And / or, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing record of the same user, and the amount of information browsed by the same user between browsing the information corresponding to the two nodes is less than a second preset threshold.

[0173] In some possible implementations of the embodiments of the present application, the text construction submodule may specifically include: an arrangement submodule and a replacement submodule;

[0174] An arrangement submodule, used for arranging the nodes in each walking path in order to obtain a first text;

[0175] The replacement submodule is used to replace the nodes in the first text with an information parameter of the corresponding information to obtain the reference text.

[0176] Optionally, the information parameter is an identity, category, source, keyword, or personalized tag.

[0177] In some possible implementations of the embodiments of the present application, when the context feature is a feature vector and there are multiple features, the second acquisition submodule may specifically include: a first splicing submodule or a second splicing submodule;

[0178] A first concatenation submodule is used to concatenate the context features of each information parameter in the reference text to obtain information features of the corresponding information;

[0179] The second concatenation submodule is used to concatenate the context features of each information parameter in the reference text by multiplying the corresponding weight to obtain the information features of the corresponding information.

[0180] In some possible implementations of the embodiments of the present application, when the context feature is a feature vector and there are multiple context features, the dimensions of the context features corresponding to different information parameters are not all the same.

[0181] In some possible implementations of the embodiment of the present application, the feature acquisition module 801 may further include:

[0182] A determination submodule, used to determine reference information in the information browsing records of multiple users that has the same information parameters as the cold information; the cold information is information that is not included in the information browsing records of multiple users;

[0183] The third acquisition submodule is used to obtain the information features of the cold information according to the information features of the reference information.

[0184] In some possible implementations of the embodiments of the present application, when the information feature is a feature vector, the feature acquisition module 801 may further include: a first calculation submodule or a second calculation submodule;

[0185] A first calculation submodule, used to take the average value of the information feature of each information in the information browsing record of the corresponding user within the first preset time period as the user feature of the corresponding user;

[0186] The second calculation submodule is used to calculate an average value of information features of each information in the information browsing record of the corresponding user within the first preset time period multiplied by a corresponding coefficient as the user feature of the corresponding user.

[0187] In some possible implementations of the embodiments of the present application, when both the information features and the user features are feature vectors, the information recommendation module 802 may specifically include:

[0188] The parameter determination submodule is used to obtain the recommendation parameter according to the distance between the user characteristics of the target user and the information characteristics of the information to be recommended; the recommendation parameter is negatively correlated with the distance;

[0189] A judging submodule, used to judge whether the recommended parameter is greater than a third preset threshold;

[0190] The recommendation submodule is used to recommend the information to be recommended to the target user when the judgment submodule judges that the recommendation parameter is greater than a third preset threshold.

[0191] In some possible implementations of the embodiments of the present application, when both the information features and the user features are feature vectors, the information recommendation module 802 may specifically include:

[0192] The parameter determination submodule is used to obtain the recommendation parameters of each information to be recommended according to the distance between the user characteristics of the target user and the information characteristics of each information to be recommended in the information set to be recommended; the recommendation parameter is negatively correlated with the distance;

[0193] The recommendation submodule is used to sort the recommendation parameters of each information to be recommended, and recommend the information to be recommended corresponding to the largest first N recommendation parameters to the target user; N is a positive integer.

[0194] In an embodiment of the present application, the information characteristics of each information can be determined based on the information browsing records of multiple users, and the user characteristics of the target user can be obtained based on the information characteristics of each information in the information browsing records of the target user within the first preset time period. Then, according to the similarity between the information characteristics of the information and the user characteristics of the target user, the information can be accurately recommended to the target user. Because the information characteristics reflect the relevance of the information with other information in the information browsing records of multiple users, and the user characteristics of the target user comprehensively reflect the information characteristics of the target user's browsing information, representing the information browsing tendency and trend of the target user, when the similarity between the information characteristics and the user characteristics is high, it means that the information corresponding to the information characteristics is consistent with the user's information browsing tendency and trend, and recommending the information to the target user can improve the accuracy of information recommendation. In addition, the information recommendation method and device provided in the embodiment of the present application are not limited by information categories or tags, and the recommended information is diverse, which improves the effect of information recommendation.

[0195] Fig. 9 1 is a block diagram of a device 900 for information recommendation according to an exemplary embodiment. For example, the device 900 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0196] Reference Fig. 9, the device 900 may include one or more of the following components: a processing component 902 , a memory 904 , a power component 906 , a multimedia component 908 , an audio component 910 , an input / output (I / O) interface 912 , a sensor component 914 , and a communication component 916 .

[0197] The processing component 902 generally controls the overall operation of the device 900, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 902 may include one or more processors 920 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 902 may include one or more modules to facilitate the interaction between the processing component 902 and other components. For example, the processing component 902 may include a multimedia module to facilitate the interaction between the multimedia component 908 and the processing component 902.

[0198] The memory 904 is configured to store various types of data to support operations on the device 900. Examples of such data include instructions for any application or method operating on the device 900, contact data, phone book data, messages, pictures, videos, etc. The memory 904 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0199] The power supply component 906 provides power to the various components of the device 900. The power supply component 906 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 900.

[0200] The multimedia component 908 includes a screen that provides an output interface between the device 900 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 908 includes a front camera and / or a rear camera. When the device 900 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0201] The audio component 910 is configured to output and / or input audio signals. For example, the audio component 910 includes a microphone (MIC), and when the device 900 is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 904 or sent via the communication component 916. In some embodiments, the audio component 910 also includes a speaker for outputting audio signals.

[0202] I / O interface 912 provides an interface between processing component 902 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.

[0203] The sensor assembly 914 includes one or more sensors for providing various aspects of status assessment for the device 900. For example, the sensor assembly 914 can detect the open / closed state of the device 900, the relative positioning of components, such as the display and keypad of the device 900, and the sensor assembly 914 can also detect the position change of the device 900 or a component of the device 900, the presence or absence of user contact with the device 900, the orientation or acceleration / deceleration of the device 900, and the temperature change of the device 900. The sensor assembly 914 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 914 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 914 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0204] The communication component 916 is configured to facilitate wired or wireless communication between the device 900 and other devices. The device 900 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 916 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 916 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0205] In an exemplary embodiment, the apparatus 900 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above methods.

[0206] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 904 including instructions, and the instructions can be executed by a processor 920 of the device 900 to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0207] A non-transitory computer-readable storage medium, when instructions in the storage medium are executed by a processor of a mobile terminal, enables the mobile terminal to execute any one of the information recommendation methods provided in the above embodiments.

[0208] Fig.101 is a schematic diagram of the structure of the server in the embodiment of the present application. The server 1000 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 1022 (for example, one or more processors) and memory 1032, and one or more storage media 1030 (for example, one or more mass storage devices) storing application programs 1042 or data 1044. Among them, the memory 1032 and the storage medium 1030 can be short-term storage or permanent storage. The program stored in the storage medium 1030 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the server. Furthermore, the central processing unit 1022 can be configured to communicate with the storage medium 1030 and execute a series of instruction operations in the storage medium 1030 on the server 1000.

[0209] The server 1000 may also include one or more power supplies 1026, one or more wired or wireless network interfaces 1050, one or more input and output interfaces 1058, one or more keyboards 1056, and / or one or more operating systems 1041, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.

[0210] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0211] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0212] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprises a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0213] The above is only a preferred embodiment of the present application and does not constitute any formal limitation to the present application. Although the present application has been disclosed as above with a preferred embodiment, it is not intended to limit the present application. Any technician familiar with the art can make many possible changes and modifications to the technical solution of the present application using the methods and technical contents disclosed above without departing from the scope of the technical solution of the present application, or modify it into an equivalent embodiment of equivalent changes. Therefore, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present application without departing from the content of the technical solution of the present application still falls within the scope of protection of the technical solution of the present application.

Claims

1. An information recommendation method, characterized in that: The method comprises: Obtain user characteristics of target users and information characteristics of information; Recommending information to the target user based on the similarity between the information feature and the user feature; The information feature includes a context feature of the corresponding information in the reference text, and the context feature is used to indicate the correlation between the target information and other information that the target user has operated when the target information is operated by the user; the reference text is obtained based on the information browsing records of multiple users; the information feature is obtained based on the correlation of information in the information browsing records of multiple users; the user feature is obtained based on the information feature of each information in the information browsing record of the corresponding user within a first preset time period; The information features are obtained by following the steps: Acquiring information browsing records of multiple users within a second preset time period; Based on the information browsing records of multiple users, an information relationship network is constructed; wherein the nodes in the information relationship network correspond to the information in the information browsing records one by one, and the line between two nodes in the information relationship network represents that the information corresponding to the two nodes appears in the information browsing records of the same user at the same time, and the information corresponding to the two nodes is operated by the same user in sequence; the line between two nodes in the information relationship network has a direction pointing to one of the nodes, and the direction of the line is determined by the order in which the same user operates the information corresponding to the two nodes connected by the line; According to preset walking parameters, performing at least one random walk in the information relationship network to obtain various walking paths; Constructing the reference text based on each obtained wandering path; Obtaining contextual features of the information corresponding to the nodes in the information relationship network in the reference text; Based on the context feature, obtaining the information feature of the corresponding information; When the information feature is a feature vector, the user feature is obtained by following the steps below: taking the average value of the information feature of each information in the information browsing record of the corresponding user within the first preset time period as the user feature of the corresponding user; Alternatively, the average value of the information feature of each information in the information browsing record of the corresponding user within the first preset time period multiplied by the corresponding coefficient is used as the user feature of the corresponding user.

2. The method according to claim 1, characterized in that The connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing record of the same user, and the time interval between the operation times of the same user on the information corresponding to the two nodes is less than the first preset threshold; And / or, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing record of the same user, and the amount of information browsed by the same user between browsing the information corresponding to the two nodes is less than a second preset threshold.

3. The method according to claim 1, characterized in that The step of constructing the reference text based on each obtained walking path specifically includes: Arrange the nodes in each of the walking paths in order to obtain a first text; The nodes in the first text are replaced with an information parameter of corresponding information to obtain the reference text.

4. The method according to claim 3, characterized in that When the context feature is a feature vector and there are multiple features, obtaining the information feature of the corresponding information based on the context feature specifically includes: Concatenating the context features of each information parameter in the reference text to obtain the information features of the corresponding information; Alternatively, the context feature of each information parameter in the reference text is multiplied by the corresponding weight and then concatenated to obtain the information feature of the corresponding information.

5. The method according to claim 3, characterized in that: When the context feature is a feature vector and there are multiple context features, the dimensions of the context features corresponding to different information parameters are not all the same.

6. The method according to claim 1, characterized in that For cold information that is not included in the information browsing records of the multiple users, information features of the cold information are obtained by using the following steps: Determine reference information in the information browsing records of the multiple users that has the same information parameters as the cold information; The information feature of the cold information is obtained according to the information feature of the reference information.

7. The method according to claim 3, characterized in that The information parameter is an identity identifier, category, source, keyword or personalized tag.

8. The method according to claim 1, characterized in that When both the information feature and the user feature are feature vectors, recommending information to the target user according to the similarity between the information feature and the user feature specifically includes: Obtaining a recommendation parameter according to the distance between the user feature of the target user and the information feature of the information to be recommended; the recommendation parameter is negatively correlated with the distance; Determining whether the recommended parameter is greater than a third preset threshold; When the recommendation parameter is greater than the third preset threshold, the information to be recommended is recommended to the target user.

9. The method according to claim 1, characterized in that: When both the information feature and the user feature are feature vectors, recommending information to the target user according to the similarity between the information feature and the user feature specifically includes: According to the distance between the user feature of the target user and the information feature of each piece of information to be recommended in the set of information to be recommended, a recommendation parameter of each piece of information to be recommended is obtained; the recommendation parameter is negatively correlated with the distance; The recommendation parameters of each of the information to be recommended are sorted, and the information to be recommended corresponding to the largest first N recommendation parameters is recommended to the target user; N is a positive integer.

10. An information recommendation device, characterized in that: The device comprises: A feature acquisition module is used to acquire user features of target users and information features of information; An information recommendation module, used to recommend information to the target user based on the similarity between the information feature and the user feature; The information feature includes a context feature of the corresponding information in the reference text, and the context feature is used to indicate the correlation between the target information and other information that the target user has operated when the target information is operated by the user; the reference text is obtained based on the information browsing records of multiple users; the information feature is obtained based on the correlation of information in the information browsing records of multiple users; the user feature is obtained based on the information feature of each information in the information browsing record of the corresponding user within a first preset time period; The feature acquisition module comprises: A record acquisition submodule, used to acquire information browsing records of multiple users within a second preset time period; The network construction submodule is used to construct an information relationship network based on the information browsing records of multiple users; wherein the nodes in the information relationship network correspond to the information in the information browsing records one by one, and the line between two nodes in the information relationship network represents that the information corresponding to the two nodes appears in the information browsing records of the same user at the same time, and the information corresponding to the two nodes is operated by the same user in sequence; the line between two nodes in the information relationship network has a direction pointing to one of the nodes, and the direction of the line is determined by the order in which the same user operates the information corresponding to the two nodes connected by the line; A random walk submodule, used to perform at least one random walk in the information relationship network according to preset walk parameters to obtain various walk paths; A text construction submodule, used for constructing the reference text based on each obtained wandering path; A first acquisition submodule is used to obtain context features of the information corresponding to the nodes in the information relationship network in the reference text; A second acquisition submodule, configured to obtain the information feature of the corresponding information based on the context feature; When the information feature is a feature vector, the feature acquisition module further includes: a first calculation submodule or a second calculation submodule; A first calculation submodule, configured to take an average value of the information feature of each information in the information browsing record of the corresponding user within the first preset time period as a user feature of the corresponding user; The second calculation submodule is used to calculate an average value of the information feature of each information in the information browsing record of the corresponding user within the first preset time period multiplied by a corresponding coefficient as the user feature of the corresponding user.

11. The device according to claim 10, characterized in that The connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing record of the same user, and the time interval between the operation times of the same user on the information corresponding to the two nodes is less than the first preset threshold; And / or, the connection between two nodes in the information relationship network represents that the information corresponding to the two nodes appears simultaneously in the information browsing record of the same user, and the amount of information browsed by the same user between browsing the information corresponding to the two nodes is less than a second preset threshold.

12. The device according to claim 10, characterized in that The text construction submodule includes: an arrangement submodule and a replacement submodule; An arrangement submodule, used for arranging the nodes in each of the walking paths in order to obtain a first text; The replacement submodule is used to replace the nodes in the first text with an information parameter of the corresponding information to obtain the reference text.

13. The device according to claim 12, characterized in that When the context feature is a feature vector and there are multiple features, the second acquisition submodule includes: a first splicing submodule or a second splicing submodule; A first splicing submodule, used for splicing the context features of each information parameter in the reference text to obtain the information features of the corresponding information; The second concatenation submodule is used to concatenate the context features of each information parameter in the reference text by multiplying the corresponding weight to obtain the information features of the corresponding information.

14. The device according to claim 10, characterized in that For cold information that is not included in the information browsing records of the multiple users, the feature acquisition module further includes: A determination submodule, used to determine reference information in the information browsing records of the plurality of users that has the same information parameters as the cold information; The third acquisition submodule is used to obtain the information features of the cold information according to the information features of the reference information.

15. The device according to claim 10, characterized in that When both the information feature and the user feature are feature vectors, the information recommendation module includes: A parameter determination submodule, used for obtaining a recommendation parameter according to the distance between the user feature of the target user and the information feature of the information to be recommended; the recommendation parameter is negatively correlated with the distance; A judging submodule, used to judge whether the recommended parameter is greater than a third preset threshold; The recommendation submodule is used to recommend the information to be recommended to the target user when the recommendation parameter is greater than the third preset threshold.

16. The device according to claim 10, characterized in that When both the information feature and the user feature are feature vectors, the information recommendation module includes: A parameter determination submodule, used for obtaining a recommendation parameter of each of the to-be-recommended information according to the distance between the user characteristic of the target user and the information characteristic of each of the to-be-recommended information in the to-be-recommended information set; the recommendation parameter is negatively correlated with the distance; The recommendation submodule is used to sort the recommendation parameters of each of the information to be recommended, and recommend the information to be recommended corresponding to the largest first N recommendation parameters to the target user; N is a positive integer.

17. A device for information recommendation, characterized in that: The device comprises a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors to perform the information recommendation method according to any one of claims 1 to 9.

18. A computer-readable medium having instructions stored thereon, which, when executed by one or more processors, causes a device to perform the information recommendation method according to any one of claims 1 to 9.

Images