Recommendation method, device, electronic device and storage medium

Abstract

The present disclosure relates to a recommendation method, device, electronic device and storage medium, including: obtaining an initial user behavior vector, wherein the initial user behavior vector lacks a first feature vector; obtaining a preset user behavior vector associated with the initial user behavior vector; based on the similarity between the initial user behavior vector and the preset user behavior vector, updating the first feature vector of the initial user behavior vector to obtain a target user behavior vector; and performing content recommendation based on the target user behavior vector to improve the accuracy of the recommendation result.

Description

Technical Field

[0001] The present disclosure relates to the technical field of search engines, and in particular to a recommendation method, device, electronic device and storage medium. Background Art

[0002] Search engines are a type of search technology that uses specific strategies to retrieve information from the Internet and feed it back to users based on user needs and certain algorithms. Search engines rely on a variety of technologies, such as web crawler technology, search and ranking technology, web page processing technology, big data processing technology, natural language processing technology, etc., to provide fast and highly relevant information services to information search users.

[0003] In the recommendation algorithm of search engines, user features are usually combined with content features for training. When user logs and content information are rich, the algorithm often has dozens or even hundreds of features.

[0004] However, not all content has all the required feature information. When features are missing, a default value is usually assigned for training. This method of defaulting to missing features will affect the training effect of the algorithm in many cases. If the data is missing important features, the algorithm will not be able to learn accurate information when the features are assigned by default, thus affecting the accuracy of the recommendation. Summary of the invention

[0005] In order to solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a recommendation method, device, electronic device and storage medium, which improve the accuracy of recommendation results.

[0006] In a first aspect, an embodiment of the present disclosure provides a recommendation method, including:

[0007] Acquire an initial user behavior vector, wherein the initial user behavior vector lacks a first feature vector;

[0008] Acquire a preset user behavior vector associated with the initial user behavior vector;

[0009] Based on the similarity between the initial user behavior vector and the preset user behavior vector, updating the first feature vector of the initial user behavior vector to obtain a target user behavior vector;

[0010] Content recommendation is performed based on the target user behavior vector.

[0011] Optionally, the updating the first feature vector of the initial user behavior vector to obtain the target user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector includes:

[0012] Extracting a second feature vector corresponding to the target feature in the preset user behavior vector and a third feature vector corresponding to the target feature in the initial user behavior vector;

[0013] Determine a first target feature vector for updating a first feature vector of the initial user behavior vector according to a similarity between a second feature vector corresponding to a target feature in the preset user behavior vector and a third feature vector corresponding to a target feature in the initial user behavior vector;

[0014] The determined first target feature vector is used to update the first feature vector in the initial user behavior vector to obtain a target user behavior vector.

[0015] Optionally, before obtaining the initial user behavior vector, the process includes:

[0016] The first feature vector is determined according to the importance of each feature vector of the user behavior vector.

[0017] Optionally, determining a first target feature vector for updating a first feature vector of an initial user behavior vector according to a similarity between a second feature vector corresponding to a target feature in the preset user behavior vector and a third feature vector corresponding to a target feature in the initial user behavior vector includes:

[0018] When the initial user behavior vector is a discrete feature, a voting method is used to determine a first target feature vector for updating the first feature vector of the initial user behavior vector according to the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector;

[0019] When the initial user behavior vector is a continuous feature, based on the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector, the first target feature vector for updating the first feature vector of the initial user behavior vector is determined by taking an average value.

[0020] Optionally, the step of using a voting method to determine a first target feature vector for updating a first feature vector of an initial user behavior vector includes:

[0021] Obtaining a first feature vector corresponding to a target feature of each of the preset user behavior vectors;

[0022] According to the first feature vectors of the preset user behavior vectors, the first feature vectors having the same number of first feature vectors among the first feature vectors of the preset user behavior vectors are used to update the first feature vector of the initial user behavior vector.

[0023] Optionally, the method of determining the first target feature vector for updating the first feature vector of the initial user behavior vector by taking the average value includes:

[0024] Obtaining a first feature vector corresponding to a target feature of each of the preset user behavior vectors;

[0025] According to the first feature vectors of the preset user behavior vectors, after averaging the first feature vectors of the preset user behavior vectors, the first target feature vector of the first feature vector of the initial user behavior vector is updated with the average value.

[0026] Optionally, the performing content recommendation based on the target user behavior vector includes:

[0027] Based on the target user behavior vector, an initial user log is updated to obtain a target user log;

[0028] Content recommendation is performed based on the target user log.

[0029] In a second aspect, an embodiment of the present disclosure provides a recommendation device, including:

[0030] An initial user behavior vector acquisition module, used to acquire an initial user behavior vector, wherein the initial user behavior vector lacks a first feature vector;

[0031] A preset user behavior vector acquisition module, used to acquire a preset user behavior vector associated with the initial user behavior vector;

[0032] A target user behavior vector determination module, configured to update a first feature vector of the initial user behavior vector to obtain a target user behavior vector based on a similarity between the initial user behavior vector and a preset user behavior vector;

[0033] A recommendation module is used to recommend content based on the target user behavior vector.

[0034] In a third aspect, an embodiment of the present disclosure provides an electronic device, including:

[0035] one or more processors;

[0036] a storage device for storing one or more programs,

[0037] When the one or more programs are executed by the one or more processors, the one or more processors implement any method as described in the first aspect.

[0038] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any method described in the first aspect.

[0039] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:

[0040] The recommendation method, device, electronic device and storage medium provided by the embodiments of the present disclosure obtain an initial user behavior vector, obtain a preset user behavior vector associated with the initial user behavior vector, update the first feature vector of the initial user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector to obtain a target user behavior vector, and recommend content based on the target user behavior vector. The target user behavior vector is obtained by complementing the missing first feature vector in the initial user behavior vector with the preset user behavior vector of the user. Therefore, the determined target user behavior vector is a complete user behavior vector, which can better represent the user behavior characteristics, and thus the content recommended based on the target user behavior vector is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0042] In order to more clearly illustrate the embodiments of the present invention 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, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0043] Figure 1 is a flowchart of a recommended method provided by an embodiment of the present disclosure;

[0044] Figure 2 is a flowchart of another recommended method provided by an embodiment of the present disclosure;

[0045] Figure 3 is a flowchart of another recommended method provided by an embodiment of the present disclosure;

[0046] Figure 4 is a flowchart of another recommended method provided by an embodiment of the present disclosure;

[0047] Figure 5 is a structural schematic diagram of a recommended device provided by an embodiment of the present disclosure;

[0048] Figure 6 It is a structural schematic diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0049] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0050] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0051] Exemplarily, the present disclosure provides a recommendation method, device, electronic device and storage medium, which obtain an initial user behavior vector, wherein the initial user behavior vector lacks a first feature vector, obtain a preset user behavior vector associated with the initial user behavior vector, update the first feature vector of the initial user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector to obtain a target user behavior vector, perform content recommendation based on the target user behavior vector, and update the first feature vector of the initial user behavior vector by analyzing the similarity between the initial user behavior vector and the preset user behavior vector, that is, complete the missing first feature vector in the initial user behavior vector, thereby improving the accuracy of search engine search results.

[0052] The recommendation method of the present disclosure is performed by an electronic device or an application, webpage, public account, etc. in the electronic device. The electronic device may be a tablet computer, mobile phone, wearable, vehicle-mounted, augmented reality (AR) / virtual reality (VR), laptop, ultra-mobile personal computer (UMPC), netbook, personal digital assistant (PDA), smart TV, smart screen, HDTV, 4K TV, smart speaker, smart projector, etc. The present disclosure does not impose any restrictions on the specific types of electronic devices.

[0053] The present disclosure does not limit the type of operating system of the electronic device, for example, Android system, Linux system, Windows system, iOS system, etc.

[0054] The technical solution of the present disclosure is described in detail below with reference to several specific embodiments.

[0055] Figure 1 A schematic diagram of a recommended method provided by the present disclosure is shown in FIG. Figure 1 As shown, the method of this embodiment is as follows:

[0056] S20: Obtain an initial user behavior vector.

[0057] Among them, the initial user behavior vector lacks the first eigenvector.

[0058] A user behavior vector refers to a vector composed of a variety of different user characteristics and / or content characteristics, which represents the user's behavior, browsing habits, browsing methods, browsing preferences, etc. when using the application or browsing the webpage. The initial user behavior vector can be, for example, the first time a user logs into a device to browse the webpage or logs into an application, or the first time a user enters a certain topic on a webpage to search.

[0059] Specifically, when a user logs into an application for the first time or browses a web page in incognito mode, the user has no user behavior vector on the logged-in application or on the web page. At this time, the first feature vector is missing in the initial user behavior vector.

[0060] S30: Acquire a preset user behavior vector associated with the initial user behavior vector.

[0061] Association refers to the same partial functions or the same partial usage scenarios, etc. When there are some identical functions and / or some identical usage scenarios in the search engine, the user behavior vector is a preset user behavior vector associated with the initial user behavior.

[0062] Exemplarily, the initial user behavior vector is obtained when a user searches based on a topic on a certain application, and the preset user behavior vector associated with the initial user behavior vector is a user behavior vector corresponding to a user searches based on the same topic on other applications.

[0063] It should be noted that, in the above embodiment, the obtained initial user behavior vector may be a user log corresponding to the same application program, or may be a user log corresponding to different application programs, and the embodiment of the present disclosure does not specifically limit this.

[0064] S40: Based on the similarity between the initial user behavior vector and the preset user behavior vector, update the first feature vector of the initial user behavior vector to obtain the target user behavior vector.

[0065] Specifically, after determining the preset user behavior vector associated with the initial user behavior vector, select the feature vector of the important dimension in the initial user behavior vector and the feature vector of the same dimension as the initial user behavior vector in the preset user behavior vector for similarity comparison, and determine the first target feature vector corresponding to the first feature vector for updating the initial user behavior vector based on the comparison result, and use the first target feature vector to update the missing first feature vector in the initial user behavior vector.

[0066] S50: Recommend content based on the target user behavior vector.

[0067] After the first feature vector of the initial user behavior vector is updated to obtain the target user behavior vector, the corresponding target user behavior vector is complemented by the user's preset user behavior vector with the missing first feature vector in the initial user behavior vector. Therefore, the determined target user behavior vector is a complete user behavior vector, which can better represent the user behavior characteristics and improve the accuracy of the recommendation results.

[0068] The recommendation method provided by the embodiment of the present disclosure obtains an initial user behavior vector, obtains a preset user behavior vector associated with the initial user behavior vector, updates the first feature vector of the initial user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector to obtain a target user behavior vector, and recommends content based on the target user behavior vector. The target user behavior vector is obtained by complementing the missing first feature vector in the initial user behavior vector with the preset user behavior vector of the user. Therefore, the determined target user behavior vector is a complete user behavior vector, which can better represent the user behavior characteristics, and thus the content recommended based on the target user behavior vector is more accurate.

[0069] Figure 2 is a flowchart of another recommended method provided by an embodiment of the present disclosure. The embodiment of the present disclosure is based on the above embodiment. Figure 2 As shown, one possible implementation of step S40 includes:

[0070] S41, extracting a second feature vector corresponding to a target feature in a preset user behavior vector and a third feature vector corresponding to a target feature in an initial user behavior vector.

[0071] Exemplarily, when the preset user behavior vector is a 100-dimensional feature vector, the initial user behavior vector is a 100-dimensional feature vector, and the 70th-dimensional feature vector in the initial user behavior vector is missing, that is, the 70th-dimensional feature vector in the initial user behavior vector is the first feature vector.

[0072] The process of selecting the second feature vector corresponding to the target feature from the preset user behavior vector and selecting the third feature vector corresponding to the target feature from the initial user behavior vector is as follows: when the target feature is the first 10-dimensional vector in the preset user behavior vector and the first 10-dimensional vector in the initial user behavior vector, the second feature vector corresponding to the target feature selected from the preset user behavior vector is the first 10-dimensional feature vector in the preset user behavior vector, and the third feature vector corresponding to the target feature selected from the initial user behavior vector is the first 10-dimensional feature vector in the initial user behavior vector.

[0073] Since the acquired preset user behavior vectors associated with the initial user behavior vector include multiple ones, the second feature vector corresponding to the target feature is extracted from the acquired multiple preset user behavior vectors in sequence.

[0074] Exemplarily, when the determined preset user behavior vectors are 50, the first 10-dimensional feature vectors in the preset user behavior vectors are extracted from the 50 preset user behavior vectors respectively.

[0075] It should be noted that, in the above embodiment, it is exemplified that the preset user behavior vector is a 100-dimensional feature vector, the initial user behavior vector is a 100-dimensional feature vector, and the target feature is the first 10-dimensional vectors in the preset user behavior vector and the first 10-dimensional vectors in the initial user behavior vector. In other possible implementations, the embodiment of the present disclosure does not specifically limit the dimensions of the preset user behavior vector and the initial user behavior vector, nor does it specifically limit the dimension of the target feature. In addition, the first 10-dimensional vectors in the preset user behavior vector are some important feature vectors of the preset user behavior vector, and the first 10-dimensional vectors in the initial user behavior vector are some important feature vectors in the initial user behavior vector. In other possible implementations, some important feature vectors of other dimensions in the preset user behavior vector may be selected, and some important feature vectors of other dimensions in the initial user behavior vector may be selected.

[0076] S42: Determine a first target feature vector for updating the first feature vector of the initial user behavior vector according to the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector.

[0077] After extracting the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector, the first target feature vector for updating the first feature vector of the initial user behavior vector is determined by comparing the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector. Specifically, the similarity between the second feature vector corresponding to the first dimension in the preset user behavior vector and the third feature vector corresponding to the first dimension in the initial user behavior vector is first compared, and then the similarity between the second feature vector corresponding to the second dimension in the preset user behavior vector and the third feature vector corresponding to the second dimension in the initial user behavior vector is compared, and so on, until the similarity comparison between the second feature vector corresponding to the tenth dimension in the preset user behavior vector and the third feature vector corresponding to the tenth dimension in the initial user behavior vector is completed, and based on the comparison result, the first target feature vector for updating the first feature vector of the initial user behavior vector is determined.

[0078] In addition, in the process of comparing the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector, the similarity between the second feature vectors corresponding to the first dimension in the multiple preset user behavior vectors and the third feature vector corresponding to the first dimension in the initial user behavior vector can be compared, and then the similarity between the second feature vectors corresponding to the second dimension in the multiple preset user behavior vectors and the third feature vector corresponding to the second dimension in the initial user behavior vector can be compared, and so on. The similarity between the second feature vector corresponding to the first dimension in the first preset user behavior vector and the third feature vector corresponding to the first dimension in the initial user behavior vector, the similarity between the second feature vector corresponding to the second dimension in the first preset user behavior vector and the third feature vector corresponding to the second dimension in the initial user behavior vector can also be compared, until the similarity between the second feature vector corresponding to the tenth dimension in the first preset user behavior vector and the third feature vector corresponding to the tenth dimension in the initial user behavior vector is compared, and then the similarity between the second feature vector corresponding to the first dimension in the second first preset user behavior vector and the third feature vector corresponding to the first dimension in the initial user behavior vector is compared. The embodiment of the present disclosure does not specifically limit the comparison process.

[0079] Optionally, the specific process of determining the first target feature vector for updating the first feature vector of the initial user behavior vector is:

[0080] When the initial user behavior vector is a discrete feature, a voting method is used to determine the first target feature vector for updating the first feature vector of the initial user behavior vector based on the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector.

[0081] Specifically, the first feature vector corresponding to the target feature of each preset user behavior vector is obtained, and according to the first feature vector of each preset user behavior vector, the first feature vector having the same number of first feature vectors among the first feature vectors of each preset user behavior vector is used to update the first feature vector of the initial user behavior vector.

[0082] Exemplarily, when the preset user behavior vector is a 100-dimensional feature vector, the initial user behavior vector is a 100-dimensional feature vector, and the 70th-dimensional feature vector in the initial user behavior vector is missing, that is, the 70th-dimensional feature vector in the initial user behavior vector is the first feature vector, 50 preset user behavior vectors associated with the initial user behavior vector are obtained, and the target feature corresponds to the first 5-dimensional feature vectors in the preset user behavior vector. When the initial user behavior vector is a discrete feature, firstly, the feature vectors corresponding to the first 5 dimensions in the 50 preset user behavior vectors are obtained respectively, and then the feature vectors corresponding to the first 5 dimensions in the initial user behavior vector are obtained. By comparing the similarity between each preset user behavior vector and the initial user behavior vector in turn, based on the similarity between each preset user behavior vector and the initial user behavior vector, the first 10 preset user behavior vectors with higher similarity are selected, and then the first target feature vector for updating the first feature vector of the initial user behavior vector is determined by a voting method based on the first 10 preset user behavior vectors with higher similarity. For example, when the feature vectors of the 70th dimension of the top 10 preset user behavior vectors with higher similarity are A1, A1, A1, A2 and A3 respectively, the first feature vectors of the first feature vectors of each preset user behavior vector with the same number of first feature vectors are used to update the first feature vector of the initial user behavior vector, that is, A1 is used to update the first feature vector of the initial user behavior vector.

[0083] When the initial user behavior vector is a continuous feature, the first target feature vector for updating the first feature vector of the initial user behavior vector is determined by taking an average value based on the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector.

[0084] Specifically, the first feature vector of each preset user behavior vector is obtained, and the first feature vectors of each preset user behavior vector are averaged according to the first feature vectors of each preset user behavior vector, and the first target feature vector of the first feature vector of the initial user behavior vector is updated with the average value.

[0085] Exemplarily, when the preset user behavior vector is a 100-dimensional feature vector, the initial user behavior vector is a 100-dimensional feature vector, the 70th-dimensional feature vector in the initial user behavior vector is missing, that is, the 70th-dimensional feature vector in the initial user behavior vector is the first feature vector, 50 preset user behavior vectors associated with the initial user behavior vector are obtained, and the target feature corresponds to the first 5-dimensional feature vector in the preset user behavior vector. When the initial user behavior vector is a continuous feature, the feature vectors corresponding to the first 5 dimensions of the 50 preset user behavior vectors are first obtained respectively, and then the feature vectors corresponding to the first 5 dimensions of the initial user behavior vector are obtained. By comparing the similarity between each preset user behavior vector and the initial user behavior vector in turn, based on the similarity between the feature vectors corresponding to the first 5 dimensions of each preset user behavior vector and the feature vectors corresponding to the first 5 dimensions of the initial user behavior vector, the first 10 preset user behavior vectors with higher similarity are selected, and then the first target feature vector for updating the first feature vector of the initial user behavior vector is determined by the averaging method based on the first 10 preset user behavior vectors with higher similarity. For example, when the feature vectors of the top 10 preset user behavior vectors with higher similarity in the 70th position are A1, A1, A1, A2 and A3 respectively, the first feature vectors of each preset user behavior vector are averaged, and the average value is used to update the first target feature vector of the first feature vector of the initial user behavior vector, that is, A1, A1, A1, A2 and A3 are averaged to update the first feature vector of the initial user behavior vector.

[0086] It should be noted that, in the above embodiment, when the initial user behavior vector obtained is a discrete feature, the preset user behavior vector obtained and associated with the initial user behavior vector is also a discrete feature; when the initial user behavior vector obtained is a continuous feature, the preset user behavior vector obtained and associated with the initial user behavior vector is also a continuous feature.

[0087] S43: Update the first feature vector in the initial user behavior vector with the determined first target feature vector to obtain a target user behavior vector.

[0088] After determining the first target feature vector of the first feature vector that updates the initial user behavior vector, the first target feature vector is used to update the first feature vector of the initial user behavior vector to obtain the target user behavior vector. At this time, the corresponding target user behavior vector is complemented by the user's preset user behavior vector with the missing first feature vector in the initial user behavior vector. Therefore, the determined target user behavior vector is a complete user behavior vector, which can better represent the user behavior characteristics and improve the accuracy of the recommendation results.

[0089] It should be noted that in the above embodiments, the target feature may be one or more, and the embodiments of the present disclosure do not specifically limit the number of the target features.

[0090] Figure 3 is a flowchart of another recommended method provided by an embodiment of the present disclosure. The embodiment of the present disclosure is based on the above embodiment. Figure 3 As shown, before step S20, the following steps are also included:

[0091] S10. Determine a first feature vector according to the importance of each feature vector of the user behavior vector.

[0092] Specifically, the search engine may be a search engine that covers all information sources (such as websites, applications, devices, etc.), or it may be an in-site search engine of any information source (such as a search function within an App or a voice search on a smart device, etc.).

[0093] Specifically, the first feature vector may include intent features, device features, scene features, behavior features, time features, theme features, download volume features, rating features, keyword features, creation time features, etc. Exemplary, intent features, such as current intent features and last intent features, device features, such as the current identification identifier and the device identifier of the last user search, scene features, such as the current scene identifier and the last user scene identifier, behavior features, such as the average length of time users stay in the current vertical domain, the average length of time users stay in the current intent, the vertical domain most clicked by users, the APP most clicked by the current user, etc., time features, such as time features: breakfast, noon, afternoon, etc., hourly level features.

[0094] The specific process of determining the first eigenvector according to the importance of each eigenvector of the user behavior vector is as follows:

[0095] When a user enters a keyword in a search engine, an initial user behavior vector is formed based on the current user's search. Since the initial user behavior vector has many feature dimensions, the present invention only selects features of important dimensions from all features as the first feature vector, and analyzes whether the first feature vector is missing in each acquired user behavior vector. When it is determined that the user behavior vector is missing the first feature vector, the current user behavior vector is determined to be the initial user behavior vector.

[0096] The recommendation method provided by the embodiment of the present disclosure determines the first feature vector according to the importance of each feature vector of the user behavior vector, that is, selects the user behavior vector that lacks the first feature vector as the initial user behavior vector, and determining the first feature vector based on the importance of each feature vector of the user behavior vector can ensure that the determined initial user behavior vector is a vector lacking important features, and then based on the subsequent update of the first feature vector of the initial user behavior vector, the accuracy of the retrieval result of the initial user behavior vector lacking important features can be guaranteed.

[0097] Figure 4 is a flowchart of another recommended method provided by an embodiment of the present disclosure. The embodiment of the present disclosure is based on the above embodiment. Figure 4 As shown, one possible implementation of step S50 includes:

[0098] S51. Update the initial user log based on the target user behavior vector to obtain the target user log.

[0099] After the first feature vector of the initial user behavior vector is updated to obtain the target user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector, the initial user log is updated to obtain the target user log based on the u target user behavior vector, so that the user feature vector of the important dimension missing from the initial user log is updated to obtain the target user log.

[0100] S52: Recommend content based on the target user log.

[0101] When a certain application or browser search is applied again in the future, content recommendations are made based on the target user log to improve the accuracy of the recommendation results.

[0102] Figure 5 is a schematic diagram of a recommended device provided by an embodiment of the present disclosure, such as Figure 5 As shown, the recommended equipment includes:

[0103] An initial user behavior vector acquisition module 510 is used to acquire an initial user behavior vector, wherein the initial user behavior vector lacks a first feature vector;

[0104] A preset user behavior vector acquisition module 520, used to acquire a preset user behavior vector associated with the initial user behavior vector;

[0105] A target user behavior vector determination module 530 is used to update the first feature vector of the initial user behavior vector to obtain a target user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector;

[0106] The recommendation module 540 is used to recommend content based on the target user behavior vector.

[0107] In the recommendation device provided by the embodiment of the present disclosure, the initial user behavior vector acquisition module acquires the initial user behavior vector, the preset user behavior vector acquisition module acquires the preset user behavior vector associated with the initial user behavior vector, the target user behavior vector determination module updates the first feature vector of the initial user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector to obtain the target user behavior vector, and the recommendation module recommends content based on the target user behavior vector. Therefore, the determined target user behavior vector is a complete user behavior vector, which can better represent the user behavior characteristics, and thus the content recommended based on the target user behavior vector is more accurate.

[0108] Optionally, the target user behavior vector determination module includes:

[0109] A feature vector extraction unit, used to extract a second feature vector corresponding to a target feature in a preset user behavior vector and a third feature vector corresponding to a target feature in an initial user behavior vector;

[0110] A first target feature vector determining unit, configured to determine a first target feature vector for updating a first feature vector of an initial user behavior vector according to a similarity between a second feature vector corresponding to a target feature in a preset user behavior vector and a third feature vector corresponding to a target feature in an initial user behavior vector;

[0111] The target user behavior vector determining unit is used to update the first feature vector in the initial user behavior vector with the determined first target feature vector to obtain the target user behavior vector.

[0112] Optionally, also include:

[0113] The first feature vector determination module is used to determine the first feature vector according to the importance of each feature vector of the user behavior vector.

[0114] Optionally, the first target feature vector determining unit includes:

[0115] A first determining unit is used to determine, when the initial user behavior vector is a discrete feature, a first target feature vector for updating the first feature vector of the initial user behavior vector by a voting method according to a similarity between a second feature vector corresponding to the target feature in the preset user behavior vector and a third feature vector corresponding to the target feature in the initial user behavior vector;

[0116] The second determination unit is used to determine the first target feature vector for updating the first feature vector of the initial user behavior vector by taking an average value based on the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector when the initial user behavior vector is a continuous feature.

[0117] Optionally, the first determining unit is specifically configured to:

[0118] Obtaining a first feature vector corresponding to a target feature of each preset user behavior vector;

[0119] According to the first feature vectors of the preset user behavior vectors, the first feature vectors having the same number of first feature vectors among the first feature vectors of the preset user behavior vectors are used to update the first feature vector of the initial user behavior vector.

[0120] Optionally, the second determining unit is specifically configured to:

[0121] Obtaining a first feature vector corresponding to a target feature of each preset user behavior vector;

[0122] According to the first feature vectors of the preset user behavior vectors, after averaging the first feature vectors of the preset user behavior vectors, the first target feature vector of the first feature vector of the initial user behavior vector is updated with the average value.

[0123] Optional, recommended modules include:

[0124] A target user log determination unit, configured to update the initial user log based on the target user behavior vector to obtain the target user log;

[0125] The recommendation unit is used to recommend content based on the target user log.

[0126] The device provided by the embodiment of the present invention can execute the method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0127] It is worth noting that in the embodiment of the above-mentioned device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0128] Figure 6 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure, such as Figure 6 As shown, the electronic device includes a processor 710, a memory 720, an input device 730, and an output device 740; the number of processors 710 in the computer device can be one or more. Figure 6 A processor 710 is taken as an example; the processor 710, the memory 720, the input device 730 and the output device 740 in the electronic device can be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0129] The memory 720 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the method in the embodiment of the present invention. The processor 710 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 720, that is, implementing the method provided in the embodiment of the present invention.

[0130] The memory 720 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 720 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 720 may further include a memory remotely arranged relative to the processor 710, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0131] The input device 730 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the electronic device, and may include a keyboard, a mouse, etc. The output device 740 may include a display device such as a display screen.

[0132] The embodiment of the present disclosure also provides a storage medium containing computer executable instructions, and the computer executable instructions are used to implement the method provided by the embodiment of the present disclosure when executed by a computer processor.

[0133] Of course, the computer executable instructions of a storage medium including computer executable instructions provided by an embodiment of the present invention are not limited to the operations of the method described above, but can also execute related operations in the method provided by any embodiment of the present invention.

[0134] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (FLASH), a hard disk or an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0135] It should be noted that, in this article, relational terms such as "first" and "second" 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. Moreover, the terms "include", "comprise" 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 sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0136] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A recommendation method, It is characterized in that include: Acquire an initial user behavior vector, wherein the initial user behavior vector lacks a first feature vector; Acquire a preset user behavior vector associated with the initial user behavior vector; Based on the similarity between the initial user behavior vector and the preset user behavior vector, updating the first feature vector of the initial user behavior vector to obtain a target user behavior vector; Performing content recommendation based on the target user behavior vector; The updating of the first feature vector of the initial user behavior vector to obtain the target user behavior vector based on the similarity between the initial user behavior vector and the preset user behavior vector includes: Extracting a second feature vector corresponding to the target feature in the preset user behavior vector and a third feature vector corresponding to the target feature in the initial user behavior vector; Determine a first target feature vector for updating a first feature vector of the initial user behavior vector according to a similarity between a second feature vector corresponding to a target feature in the preset user behavior vector and a third feature vector corresponding to a target feature in the initial user behavior vector; The determined first target feature vector is used to update the first feature vector in the initial user behavior vector to obtain a target user behavior vector; The determining, based on the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector, a first target feature vector for updating the first feature vector of the initial user behavior vector comprises: When the initial user behavior vector is a discrete feature, a voting method is used to determine a first target feature vector for updating the first feature vector of the initial user behavior vector according to the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector; When the initial user behavior vector is a continuous feature, based on the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector, the first target feature vector for updating the first feature vector of the initial user behavior vector is determined by taking an average value.

2. The method according to claim 1, It is characterized in that Before obtaining the initial user behavior vector, the following steps are included: The first feature vector is determined according to the importance of each feature vector of the user behavior vector.

3. The method according to claim 1, It is characterized in that The method of using a voting method to determine a first target feature vector for updating a first feature vector of an initial user behavior vector includes: Obtaining a first feature vector corresponding to a target feature of each of the preset user behavior vectors; According to the first feature vectors of the preset user behavior vectors, the first feature vectors having the same number of first feature vectors among the first feature vectors of the preset user behavior vectors are used to update the first feature vector of the initial user behavior vector.

4. The method according to claim 1, It is characterized in that The method of determining the first target feature vector of the first feature vector for updating the initial user behavior vector by taking the average value includes: Obtaining a first feature vector corresponding to a target feature of each of the preset user behavior vectors; According to the first feature vectors of the preset user behavior vectors, after averaging the first feature vectors of the preset user behavior vectors, the first target feature vector of the first feature vector of the initial user behavior vector is updated with the average value.

5. The method according to claim 1, It is characterized in that The performing content recommendation based on the target user behavior vector includes: Based on the target user behavior vector, an initial user log is updated to obtain a target user log; Content recommendation is performed based on the target user log.

6. A recommended device, It is characterized in that include: An initial user behavior vector acquisition module, used to acquire an initial user behavior vector, wherein the initial user behavior vector lacks a first feature vector; A preset user behavior vector acquisition module, used to acquire a preset user behavior vector associated with the initial user behavior vector; A target user behavior vector determination module, configured to update a first feature vector of the initial user behavior vector to obtain a target user behavior vector based on a similarity between the initial user behavior vector and a preset user behavior vector; A recommendation module, used for recommending content based on the target user behavior vector; The target user behavior vector determination module includes: A feature vector extraction unit, used to extract a second feature vector corresponding to the target feature in the preset user behavior vector and a third feature vector corresponding to the target feature in the initial user behavior vector; A first target feature vector determining unit, configured to determine a first target feature vector for updating a first feature vector of an initial user behavior vector according to a similarity between a second feature vector corresponding to a target feature in the preset user behavior vector and a third feature vector corresponding to a target feature in the initial user behavior vector; A target user behavior vector determining unit, configured to update the first feature vector in the initial user behavior vector with the determined first target feature vector to obtain a target user behavior vector; The first target feature vector determining unit comprises: A first determining unit is configured to determine, when the initial user behavior vector is a discrete feature, a first target feature vector for updating the first feature vector of the initial user behavior vector by a voting method according to a similarity between a second feature vector corresponding to the target feature in the preset user behavior vector and a third feature vector corresponding to the target feature in the initial user behavior vector; A second determination unit is used to determine a first target feature vector for updating the first feature vector of the initial user behavior vector by taking an average value based on the similarity between the second feature vector corresponding to the target feature in the preset user behavior vector and the third feature vector corresponding to the target feature in the initial user behavior vector when the initial user behavior vector is a continuous feature.

7. An electronic device, It is characterized in that include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images