A method, device, electronic device and storage medium for recommending target information

Abstract

The present disclosure relates to a method for recommending target information, comprising: dividing request information of multiple different users into M parts, the Nth part including multiple Nth request information sets, the N+1th part including at least one N+1th request information set, the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set; N is greater than or equal to 1 and less than or equal to M-1, and N and M are both positive integers; according to the Nth weight of at least one dimension information of the information to be recommended, determining the Nth target information matching the request information in the Nth request information set from the information to be recommended; obtaining feedback information after the Nth target information is recommended; according to the feedback information, updating the N+1th weight of at least one dimension information of the information to be recommended; based on the updated N+1th weight, determining the N+1th target information matching the request information in the N+1th request information set from the information to be recommended.

Description

Technical Field

[0001] The present disclosure relates to the field of information processing, and in particular to a method, device, electronic device and storage medium for recommending target information. Background Art

[0002] With the development of science and technology, many new application technologies and application scenarios have emerged. These application technologies can be applied to relevant application scenarios. In different application scenarios, information corresponding to the application scenario will be generated. By processing this information, this information can be better applied in relevant application scenarios and achieve better application results.

[0003] For example, when recommending information, after the relevant information is processed through information processing technology, more suitable information can be recommended to the user. Summary of the invention

[0004] The present disclosure provides a method, device, electronic device and storage medium for recommending target information.

[0005] According to a first aspect of an embodiment of the present disclosure, a method for recommending target information is provided, comprising: dividing request information of multiple different users into M parts, the Nth part comprising multiple Nth request information sets, the N+1th part comprising at least one N+1th request information set, the number of request information included in the Nth request information set being less than the number of request information included in the N+1th request information set; wherein N is greater than or equal to 1 and less than or equal to M-1, and N and M are both positive integers; determining, from the information to be recommended, the Nth target information matching the request information in the Nth request information set according to the Nth weight of at least one dimension information of the information to be recommended; obtaining feedback information after the Nth target information is recommended; according to the feedback information, updating the N+1th weight of at least one dimension information of the information to be recommended; and determining, from the information to be recommended, the N+1th target information matching the request information in the N+1th request information set based on the updated N+1th weight.

[0006] In one embodiment, the Mth part includes: an Mth request information set; the method also includes: after updating the Mth weight, updating the first weights corresponding to multiple first request information sets included in the first part according to the updated Mth weight; and re-determining the first target information that matches the request information in the first request information set from the information to be recommended according to the updated first weight.

[0007] In one embodiment, after updating the Mth weight, the first weights corresponding to the multiple first request information sets included in the first part are updated according to the updated Mth weight, including: determining the same number of target weights as the first request information sets included in the first part according to the updated Mth weight; and updating the first weights according to the target weights.

[0008] In one embodiment, after updating the first weights corresponding to the multiple first request information sets included in the first part, it also includes: updating the second weight to the Mth weight based on the updated first weight; wherein the N+1th weight is updated based on the updated Nth weight.

[0009] In one embodiment, multiple request information in the Nth request information set have the same Nth weight; obtaining feedback information after the Nth target information is recommended includes: determining feedback dimension information of the Nth target information and parameters of each feedback dimension information; determining a feedback value based on the feedback dimension information and parameters of each feedback dimension information.

[0010] In one embodiment, updating the N+1th weight of at least one dimension information of the information to be recommended based on the feedback information includes: determining the feedback values ​​corresponding to the request information in multiple Nth request information sets respectively; determining the Nth weight corresponding to the request information in the Nth request information set with the largest feedback value as the Nth target weight; and updating the N+1th weight based on the Nth target weight.

[0011] In one embodiment, the feedback dimension information includes at least one of the following: download volume, collection volume, like volume, number of users following the author of the recommended information, average browsing time, number of comments, and number of reposts.

[0012] In one embodiment, the dividing of the request information of multiple different users into M parts includes: dividing the request information with the same reception time within the same time period into the same part according to the reception time of the request information; and / or obtaining a sorting value according to the priority of the user sending the request information and the sorting factor corresponding to the reception time of the request information; and dividing the request information with the sorting value within the same interval into the same part.

[0013] In one embodiment, the information to be recommended includes at least one of the following: video; public account; article; advertisement; the dimensional information includes at least one of the following: download volume, collection volume, like volume, number of users following the author of the information to be recommended, average browsing time, number of comments and number of reposts.

[0014] According to a second aspect of an embodiment of the present disclosure, there is provided a device for recommending target information, comprising: a division module, for dividing request information of multiple different users into M parts, the Nth part comprising multiple Nth request information sets, the N+1th part comprising at least one N+1th request information set, the number of request information included in the Nth request information set being less than the number of request information included in the N+1th request information set; wherein N is greater than or equal to 1 and less than or equal to M-1, and N and M are both positive integers; an Nth target information determination module, for determining, from the information to be recommended, the Nth target information that matches the request information in the Nth request information set according to the Nth weight of at least one dimension information of the information to be recommended; a feedback information acquisition module, for acquiring feedback information after the Nth target information is recommended; an updating module, for updating, according to the feedback information, the N+1th weight of at least one dimension information of the information to be recommended; an N+1th target information determination module, for determining, from the information to be recommended, the N+1th target information that matches the request information in the N+1th request information set based on the updated N+1th weight.

[0015] According to a third aspect of the present disclosure, an electronic device is provided, including:

[0016] A processor and a memory for storing executable instructions capable of running on the processor, wherein:

[0017] When the processor is used to run the executable instructions, the executable instructions execute the method described in any one of the above embodiments.

[0018] According to a fourth aspect of the embodiments of the present disclosure, a non-temporary computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, the method described in any of the above embodiments is implemented.

[0019] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects:

[0020] The disclosed embodiment first divides the request information of multiple different users into M parts, the Nth part includes multiple Nth request information sets, the N+1th part includes at least one N+1th request information set, and the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set. N is greater than or equal to 1 and less than or equal to M-1, and N and M are both positive integers. In this way, the classification of the request information is achieved, and it is divided into sets including different request information.

[0021] When determining the target information of the request information included in each request information set, each request information set has a corresponding weight of at least one dimension of the information to be recommended, for example, the Nth weight is used to determine the Nth target information of the request information in the Nth request information set. According to the Nth weight of at least one dimension of the information to be recommended, the Nth target information matching the request information in the Nth request information set is determined from the information to be recommended. In this way, the target information of the request information in each Nth request information set in the Nth part is determined.

[0022] Then, feedback information after the Nth target information is recommended is obtained, and the N+1th weight of at least one dimension of the information to be recommended is updated according to the feedback information. Based on the updated N+1th weight, the N+1th target information matching the request information in the N+1th request information set is determined from the information to be recommended. In this way, the update of the N+1th weight according to the Nth weight is realized, and the N+1th target information is determined according to the Nth weight, thereby realizing the update of the second weight according to the first weight, the update of the third weight according to the second weight, and so on, and the automatic update from the previous weight to the next weight is realized.

[0023] Since the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set, the weight of the N+1th request information set including a larger number of request information is updated according to the corresponding feedback information of the Nth request information set including a smaller number of request information, so that the determined N+1th target information is more accurate. That is, according to the feedback effect of the recommendation with small traffic, the recommendation with large traffic is updated, and the dynamic update recommendation with less data volume is realized. After updating the corresponding weight, the corresponding target information can be determined according to the updated weight, so that the determined target information has a higher matching degree with the user's request information, the recommended target information is more accurate, and a more accurate recommendation is achieved.

[0024] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] 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.

[0026] Figure 1 is a flowchart of a method for recommending target information according to an exemplary embodiment;

[0027] Figure 2 is a schematic diagram of a process of obtaining feedback information according to an exemplary embodiment;

[0028] Figure 3 is a schematic diagram of a process of updating the N+1th weight according to an exemplary embodiment;

[0029] Figure 4 Another method for recommending target information is shown according to an exemplary embodiment;

[0030] Figure 5 is a schematic structural diagram of a device for recommending target information according to an exemplary embodiment;

[0031] Figure 6 is a schematic diagram showing another recommendation method according to an exemplary embodiment;

[0032] Figure 7 The diagram is a schematic diagram of a terminal device according to an exemplary embodiment. DETAILED DESCRIPTION

[0033] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices consistent with some aspects of the present disclosure as detailed in the appended claims.

[0034] Generally, when recommending information, there are many reference factors, that is, many factors that affect the information recommendation results. When recommending information, multiple reference factors will be combined to recommend the final information. However, in general, the weights corresponding to each reference factor may be different. The final information to be recommended is determined based on the weights of these different reference factors when recommending the final information.

[0035] With the increase of reference factors, and different reference factors have different weights for different users, and the weights corresponding to these reference factors can also change due to factors such as user preferences, so it is very important to adjust the weights corresponding to these reference factors. Usually, the weight adjustment is done through offline adjustment methods such as manual adjustment. After the offline adjustment is completed, it is updated online, and it cannot be completely separated from human intervention to achieve online automatic parameter adjustment. In addition, offline adjustment methods such as manual adjustment need to rely on human experience. When there are many reference factors, it becomes increasingly difficult to adjust the parameters based on human experience.

[0036] In such a case, since it becomes increasingly difficult to adjust the weights, the weights may not be updated in a timely manner, which will lead to a decrease in the accuracy of the final recommended information and the inability to recommend the most suitable or most matching recommendation information, thereby reducing the recommendation effect and user experience.

[0037] refer to Figure 1 , is a flow chart of a method for recommending target information provided by an embodiment of the present disclosure, the data processing method comprising the following steps:

[0038] Step S100, dividing the request information of multiple different users into M parts, the Nth part includes multiple Nth request information sets, the N+1th part includes at least one N+1th request information set, and the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set. N is greater than or equal to 1 and less than or equal to M-1, and both N and M are positive integers.

[0039] Step S200: determining, from the information to be recommended, Nth target information matching the request information in the Nth request information set according to the Nth weight of at least one dimension information of the information to be recommended.

[0040] Step S300, obtaining feedback information after the Nth target information is recommended.

[0041] Step S400: updating the N+1th weight of at least one dimension of information to be recommended according to the feedback information.

[0042] Step S500: Based on the updated N+1th weight, determine the N+1th target information matching the request information in the N+1th request information set from the information to be recommended.

[0043] The method can be executed at least in a mobile terminal, that is, the execution subject of the method can at least include a mobile terminal. Mobile terminals can include mobile phones, tablet computers, vehicle-mounted central control devices, wearable devices, smart devices, etc., and smart devices can include smart office equipment and smart home equipment, etc.

[0044] The method of this embodiment can be applied to various recommendation scenarios, which can be recommendation scenarios where the recommendation results are determined by one dimensional information, or recommendation scenarios where the recommendation results are determined by multiple dimensional information. In these recommendation scenarios, the candidate information to be recommended has multiple dimensional information that can be referenced, and the final recommendation information can be determined based on these dimensional information. Different dimensional information may have different proportions in determining the final target information, so different dimensional information may have different reference values ​​for determining the final target information, that is, the weights of different dimensional information may be different.

[0045] Of course, any scenario in which the target information to be recommended is finally determined from the information to be recommended based on these different dimensional information is within the protection scope of this embodiment.

[0046] In this embodiment, the information to be recommended includes at least one of the following: videos, public accounts, advertisements, and articles, etc. Of course, other types of content may also be included, which are not listed here one by one. As long as the content includes multiple dimensional information, it can be used as the information to be recommended and is within the protection scope of this embodiment.

[0047] In one embodiment, the dimension information of the information to be recommended includes at least one of the following: download volume, collection volume, like volume, number of users following the author of the candidate information, average browsing time, number of comments, and number of reposts, etc. Of course, different candidate information may have at least one dimension information matching the candidate information. In this embodiment, the dimension information of the information to be recommended may be generated based on the historical operations of the user who has received the information on the information to be recommended.

[0048] In another embodiment, the dimensional information of the information to be recommended includes at least one of the following: the difference between the amount of information of the information to be recommended and the amount of information preferred by the recommended user, the similarity between the information content of the information to be recommended and the content preferred by the recommended user, the strength of the information to be recommended, etc.

[0049] In specific application scenarios, for example, when watching or browsing videos, public accounts, advertisements, and articles, the videos, public accounts, advertisements, and articles may be downloaded, collected, liked, commented on, and / or forwarded, and the authors of these videos, public accounts, advertisements, or articles may also be followed, etc. Different dimensional information such as the number of downloads, the number of collections, the number of likes, the number of users who follow the author of the alternative information, the average browsing time, the number of comments, and the number of forwardings can reflect the user's preference for the videos, public accounts, advertisements, and articles that have been watched or browsed.

[0050] Different dimensional information has different weights in the information to be recommended. The weights of these dimensional information in the information to be recommended can determine which information to be recommended will be determined as target information, that is, the target information can be determined from the information to be recommended based on the weights of these dimensional information in the recommended information. When recommending subsequent videos, public accounts, advertisements or articles to users, the target information to be recommended can be determined from the information to be recommended based on the download, collection, like, comment and / or forwarding information of the video, public account, article or other form of content to be recommended.

[0051] Therefore, the weights of different dimensional information of the information to be recommended play a key role in determining the target information. This embodiment is to automatically adjust the weights of different dimensional information of the information to be recommended, thereby determining the description of the target information from the information to be recommended based on the adjusted weights.

[0052] The user's request information may include: the user's request information for recommending target information, which may be request information matching an actual application program. The specific format may be determined according to the actual application program and is not limited here.

[0053] For example, when using an application to watch a video, when switching from the currently watched video to the next video, a request message for watching the next video is generated and sent to the terminal or server. Each switching action corresponds to a request message, which is used to request the terminal or server for the next video that the current user needs to watch. The terminal or server recommends the next video to the current user based on the request message.

[0054] Of course, when the information to be recommended is other types of information, when the current content switches to the next content, request information will also be generated and sent to the terminal or the server.

[0055] Regarding step S100, when this embodiment is directed to request information from multiple different users, the multiple request information is processed, and then information is recommended for the multiple request information.

[0056] When there are request information of multiple different users, the request information of the multiple different users is divided into several parts, for example, into M parts. Then the request information in each part is divided again, for example, the first part includes ten request information sets, and each request information set includes multiple request information.

[0057] The request information set included in the Nth part of the M part is taken as the Nth request information set, and multiple Nth request information sets may be the same or different. For example, the ten request information sets in the first part are all taken as the first request information set, which can be specifically divided into the first request information set 1, the first request information set 2, the first request information set 3, ... the first request information set 10, and the number of request sets included in different first request sets may be the same or different. The five request information sets in the second part are all taken as the second request information set, which can be specifically divided into the second request information set 1, the second request information set 2, the second request information set 3, ... the second request information set 5. The request information sets included in other parts are also divided in the above manner.

[0058] The principle of division may be that the quantity of request information in the request information sets included in different parts is different, and the request information is divided in a gradually increasing or gradually decreasing order.

[0059] For example, the request information is divided into three parts, namely the first part, the second part and the third part, M is equal to 3, the first part includes five first request information sets, namely the first request set 1, the first request information set 2, the first request information set 3, the first request information set 4 and the first request information set 5, and each first request information set includes the same number of request information. At this time, N is equal to 1, and any one of the first request set 1 to the first request set 5 can be used as the first request information. The second part includes two second request information sets, namely the second request information set 1 and the second request information set 2, and each second request information set includes the same number of request information. At this time, N is equal to 2, and any one of the second request set 1 and the second request set 2 can be used as the second request information. The third part includes a third request information set.

[0060] The number of request information included in the first request information set is less than the number of request information included in the second request information set; the number of request information included in the second request information set is less than the number of request information included in the third request information set. The number of request information included in the first request information set to the number of request information included in the third request information set increases in sequence, or the number of request information included in the third request information set to the number of request information included in the first request information set decreases in sequence.

[0061] That is, for request information sets in different parts, the number of request information included in the request information combination is different, and they can be sorted in an ascending or descending manner, for example, the number of request information included in the first request information set to the number of request information included in the Mth request information set increases in sequence, or the number of request information included in the Mth request information set to the number of request information included in the first request information set decreases in sequence. The number of request information included in multiple request information sets in the same part can be the same.

[0062] For example, the number of request information included in the five first request information sets is the same, and the number of request information included in the two second request information sets is the same.

[0063] The value of N can be any value in M-1, N is greater than or equal to 1 and less than or equal to M-1, and both N and M are positive integers. For example, when N is equal to 1, the first request information set is the first request information set 1 to the first request information set 5. The N+1th request information set is the second request information set in the second part, such as the second request information set 1 and the second request information set 2. For any N, the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set.

[0064] The number of M can be determined according to the number of request information, and the number of request information sets included in each part can also be determined according to actual needs.

[0065] In one embodiment, the number of request information sets included in the Nth part is greater than the number of request information sets included in the N+1th part, that is, the number of the Nth request information sets is greater than the number of the N+1th request information sets.

[0066] In another embodiment, N is greater than or equal to 2 and less than or equal to M.

[0067] For step S200, after the request information is divided, each request information set in each part has a weight of at least one dimension of the information to be recommended, and multiple request information in the same request information set share the same weight. Different request information sets may have the same or different weights. Different request information sets in the same part may have the same or different weights; request information sets in different parts may have the same or different weights.

[0068] In this embodiment, the weight possessed by each request information set in each part may be a weight set corresponding to at least one dimension information, and the weight set includes the weight of each dimension information when determining the target information from the information to be recommended.

[0069] For example, if the dimensional information of the information to be recommended includes the seven dimensional information of download volume, collection volume, like volume, number of users followed by the author of the candidate information, average browsing time, number of comments, and number of reposts, then the weight of a certain request information set can be expressed as X = {a1, b1, c1, d1, e1, f1, g1}, and different request information sets may have different a1 to g1 in X. a1, b1, c1, d4, e1, f1, and g1 respectively represent the weight of the target information in determining the download volume, collection volume, like volume, number of users followed by the author of the candidate information, average browsing time, number of comments, and number of reposts from the information to be recommended.

[0070] For another example, the weights of the request information sets in the Nth part are taken as the Nth weights. The first request information set 1 in the first part has a first weight X1 = {a1, b1, c1, d1, e1, f1, g1}, the first request information set 2 has a first weight X2 = {a2, b2, c2, d2, e2, f2, g2}, ... The first request information set 10 has a first weight X10 = {a10, b10, c10, d10, e10, f10, g10}. The second request information set 1 in the second part has a second weight X11 = {a11, b11, c11, d11, e11, f11, g11}, and the second request information set 2 has a second weight X12 = {a12, b12, c12, d12, e12, f12, g12}. The third request information set in the third part has a third weight X13 = {a13, b13, c13, d13, e13, f13, g13}.

[0071] The weight of each request information set in each part defaults to an initial value in the initial state. The initial value can be determined based on historical weights, or randomly assigned, or determined based on the user's registration information, such as gender, occupation, and hobbies, or determined based on the user's geographic location, etc.

[0072] Then, according to the Nth weight possessed by the Nth request information set, an Nth target matching the request information in the Nth request information set is determined from the information to be recommended.

[0073] For example, the first request information set 1 in the first part has a first weight X1, and the target information matching the first request information set 1 can be determined from the information to be recommended according to the first weight X1, and the target information matching the first request information set 2 can be determined from the information to be recommended according to the first weight X2, and so on. In this way, the corresponding target information can be determined according to the weight of each request information set, and the target information matching each request information set in each part can be determined, so as to recommend the target information to the request information in each request information set in each part.

[0074] For step S300, after determining the target information matched by the Nth request information set in the Nth part, the corresponding target information is recommended to the user corresponding to the request information in the Nth request information set. Then the corresponding user can obtain the corresponding target information and obtain the user's feedback information on the recommended target information. The feedback information may include the user's interest in the target information, so as to facilitate updating the corresponding weight according to the feedback information. The feedback information is determined by browsing time, whether to like, whether to collect, whether to forward, etc. The feedback information can be directly obtained through the terminal, and the feedback information can be presented in the form of numerical information.

[0075] Each request information set in each part has corresponding feedback information that matches the target information.

[0076] For step S400, after the feedback information is obtained, the N+1th weight of at least one dimension of the information to be recommended is updated according to the feedback information.

[0077] The feedback information of obtaining the Nth target information matching the Nth request information set in the Nth part is recorded as the Nth feedback information, and the N+1th weight is updated according to the Nth feedback information.

[0078] The main goal of this step is to determine the accuracy of the Nth weight for the recommended Nth target information through the feedback information of the Nth target determined by the Nth weight, so as to update the N+1th weight according to the Nth weight, and then realize the update of the N+1th weight.

[0079] Through the above steps, it is achieved that according to the first weight of the first request information set in the first part, the second weight of the second request information set in the second part is updated, and the weight of the third request information set in the third part is updated according to the second weight of the second request information set in the second part, and so on, updating to the Mth weight of the Mth request information set in the Mth part, thereby updating all the weights from the second request information set in the second part to the Mth weight of the Mth request information set in the Mth part.

[0080] Updating the second weight to the Mth weight can be automatically implemented without human intervention, and can be performed online to adjust the weight online. More accurate target information can be determined based on the updated weight, thereby improving the accuracy of the recommended target information, and recommending more matching target information to the user, thereby improving the user experience.

[0081] For step S500, after updating the N+1th weight, based on the updated N+1th weight, the N+1th target information that matches the request information in the N+1th request information set is determined from the information to be recommended, thereby recommending the N+1th target information to the request information in the N+1th request information set.

[0082] After updating the weights of the request information sets in each part, corresponding target information is recommended to the request information included in each request information set in each part according to the updated weights of the request information sets in each part.

[0083] Through this method, the weight corresponding to the request information set including a small number of request information is realized, and the weight corresponding to the request information set including a large number of request information is updated. First, according to the weight corresponding to the request information set including a small number of request information, the target information matching the request information set including a small number of request information is determined, and then according to the feedback information, the weight corresponding to the request information set including a large number of request information is further determined to be updated. This method saves manual participation and reduces the difficulty of manually updating weights.

[0084] On another level, according to the weight corresponding to the request information set including a smaller number of request information, when a problem occurs in the target information matching the request information set including a smaller number of request information, it will not affect a large number of users, thereby reducing the impact on users.

[0085] In another embodiment, multiple request information in the Nth request information set have the same Nth weight. Each request information set in each part includes multiple request information, and multiple request information in the same request information set have the same weight.

[0086] For example, among the ten first request information sets included in the second part, if the first request information set 1 has a first weight of 1, then multiple request information in the first request information set 1 share the first weight of 1, that is, multiple request information in the first request information set 1 have the same weight, which is 1. If the first request information set 2 has a first weight of 2, then multiple request information in the first request information set 2 share the first weight of 2, that is, multiple request information in the first request information set 2 have the same weight, which is 2. If the second request information set 1 in the second part has a second weight of 1, then multiple request information in the second request information set 1 share the second weight of 1, that is, multiple request information in the second request information set 1 have the same weight, which is 1.

[0087] For the same request information set, target information recommended to each request information in the request information set is determined from the information to be recommended according to the same weight.

[0088] refer to Figure 2 , is a flow chart of obtaining feedback information. Step S300, obtaining feedback information after the Nth target information is recommended, including:

[0089] Step S301, determining feedback dimension information of the Nth target information and parameters of each feedback dimension information.

[0090] Step S302: determining a feedback value according to the feedback dimension information and parameters of each feedback dimension information.

[0091] The main purpose of this embodiment is to determine the feedback information of the recommended target information after determining the target information recommended to the request information in the corresponding request information set according to the weights possessed by each request information set in each part, so as to determine whether the weights used in determining the target information are accurate.

[0092] For step S301, the specific feedback dimension information can be determined based on at least one dimension information of the information to be recommended, and can be the same information as at least one dimension information of the information to be recommended, that is, at least one dimension information of the information to be recommended is also used as feedback dimension information.

[0093] For example, the feedback dimension information may include at least one of the following:

[0094] Downloads, favorites, likes, number of users following the author of the recommended information, average browsing time, number of comments, and number of reposts. Of course, the corresponding feedback dimension information is different depending on the information to be recommended. The feedback dimension information listed here is the feedback dimension information when the recommended information is a video, public account, advertisement, or article.

[0095] For another example, the feedback dimension information may also include: content duration, average playback time per person, average playback times per person, etc.

[0096] Of course, the feedback dimension information may also be determined according to actual needs. For example, the feedback dimension information may be preset according to the information to be recommended, that is, the feedback dimension information is preset according to the information to be recommended.

[0097] In one embodiment, the feedback dimension information may be known.

[0098] After determining the reverse dimension information, since different feedback dimension information can reflect different feedbacks of users on the recommended target information, the parameters of each feedback dimension information are then determined. The parameters of the feedback dimension information can represent the proportion or weight of the feedback dimension information. According to the parameters of the feedback dimension information, the final feedback information result can be easily determined.

[0099] For step S302, after the feedback information and the corresponding parameters are determined, the feedback value is determined according to the feedback dimension information and the parameters of each feedback dimension information.

[0100] The goal of this step is to quantify the feedback information and present it in numerical form, so that the feedback information of target information corresponding to different request information sets can be better compared.

[0101] The feedback value can be determined based on each feedback dimension information and the sum of the parameters of each feedback dimension information. For example, the feedback dimension information includes: content duration, average playback duration per person and average playback times per person, the parameter of content duration is x, the parameter of average playback duration per person is y, and the parameter of average playback times per person is z. In the feedback information corresponding to different target information, the values ​​of the parameters of the feedback dimension information may be different. For example, in the first feedback information 1 of the target information matched by the first request information set 1, the parameters of each feedback dimension information are (x1, y1, z1), in the first feedback information 2 of the target information matched by the first request information set 2, the parameters of each feedback dimension information are (x2, y2, z2), in the second feedback information 1 of the target information matched by the second request information set 1, the parameters of each feedback dimension information are (x3, y3, z3), and so on.

[0102] Each piece of feedback information itself may have a quantified value, which may be the value of each feedback dimension information itself, or a converted value, etc. For example, the content duration, average playing duration per person, and average number of plays per person may all be specific values, and the feedback value may be determined based on the specific values ​​of the content duration, average playing duration per person, and average number of plays per person and their corresponding parameters, and the feedback value is the final result of the feedback information.

[0103] The feedback value is recorded as R, and R can be determined according to the value of each feedback dimension information and the parameter corresponding to the value of each feedback dimension information. The expression of R can be R = (x*content duration + y*average playing time + z*average number of plays).

[0104] The feedback value of each target information can be determined, so that the request information set can be matched with the feedback information, and further the weight of the request information set can be matched with the feedback information. After the feedback value of each target information is determined, the weight can be easily updated.

[0105] In another embodiment, reference Figure 3 , is a flow chart of updating the N+1th weight. Step S400, updating the N+1th weight of at least one dimension of the information to be recommended according to the feedback information, including:

[0106] Step S401: Determine feedback values ​​corresponding to request information in a plurality of Nth request information sets.

[0107] Step S402: determine the Nth weight corresponding to the request information in the Nth request information set with the largest feedback value as the Nth target weight.

[0108] Step S403, updating the N+1th weight according to the Nth target weight.

[0109] Since step S300 has determined the feedback information after the Nth target information is recommended, the feedback information corresponding to the target information matched by each request information set can be determined according to step S300. After determining the feedback information of the target information matched by each request information set in each part, the weights of the multiple N+1th request information sets in the N+1th part can be updated according to the feedback information of the target information matched by the multiple Nth request information sets in the Nth part and the weights of the multiple Nth request information sets.

[0110] Specifically, the weight of the Nth request information set with the largest corresponding feedback value can be determined from the weights corresponding to at least two Nth request information sets, and the weight can be determined as the Nth target weight. The Nth target weight is used to update the N+1th weight possessed by the N+1th request information set in the N+1th part. The larger the feedback value, the better the recommendation effect of the target information determined according to the corresponding weight, that is, the Nth weight corresponding to the target information with the best recommendation effect is used to update the N+1th weight. In this way, the N+1th target information determined according to the N+1th weight can be more accurate and more in line with the needs of users.

[0111] For example, the second request information set 1 in the second part is updated according to the feedback information of the target information matched by the first request information set 1, the first request information set 2, the first request information set 3, the first request information set 4 and the first request information set 5 in the first part. The first request information set with the largest feedback value is determined according to the feedback values ​​corresponding to the first five first request information sets, and then the first weight of the first request information set with the largest feedback value is updated as the weight of the second request information set 1 in the second part.

[0112] The feedback information of the target information matched by the other five first request information sets in the first part: first request information set 6, first request information set 7, first request information set 8, first request information set 9 and first request information set 10 are used to update another second request information set in the second part: second request information set 2. According to the feedback values ​​corresponding to the first request information sets 6 to 10, the first request information set with the largest feedback value is determined, and then the first weight of the first request information set with the largest feedback value is updated to the weight of the second request information set 2 in the second part. In this way, the second weight is updated according to the first weight.

[0113] For another example, based on the feedback values ​​corresponding to the second request information set 1 and the second request information set 2 respectively, the second request information set corresponding to the maximum feedback value is determined, and the weight of the second request information set is used to determine the parameters of the third request information set in the third part, thereby updating the third weight according to the second weight in the second part.

[0114] When the request information is divided into M parts, according to this principle, the weight of the second request information set in the adjacent second part can be updated according to the weight of the first request information set in the first part, and then the weight of the third request information set in the third part can be updated according to the updated weight of the second request information set, and so on, until the weight of the Mth request information set in the Mth part is updated, completing the update of the second weight to the Mth weight.

[0115] In another embodiment, the Mth part includes: an Mth request information set.

[0116] After the request information of multiple different users is divided into multiple parts, the part containing the largest number of request information in the request information set has only one request information set.

[0117] For example, it is divided into M parts, namely the first part to the Mth part. If M is equal to 10, the 10th part only includes one 10th request information set, and the number of request information included in the 10th request information set is the largest among all request information sets.

[0118] refer to Figure 4 , is another method for recommending target information, the method further comprising:

[0119] Step S600: After the Mth weight is updated, the first weights corresponding to the multiple first request information sets included in the first part are updated according to the updated Mth weight.

[0120] Step S700: re-determine, from the information to be recommended, first target information that matches the request information in the first request information set according to the updated first weight.

[0121] For step S600, after updating the weight of the Mth request information set in the Mth part, i.e., the Mth weight, the weight of the first request information set in the first part is updated according to the Mth weight. In this way, the second weight of the second request information set in the second part is updated according to the first weight of the first request information set in the first part, and the third weight of the third request information set in the third part is updated according to the updated second weight, and so on, until the Mth weight of the Mth request information set in the Mth part is updated according to the updated M-1th weight of the M-1th request information set in the M-1th part, and then the first weight is updated according to the updated Mth weight, thus forming a weight update cycle.

[0122] The specific update process is not limited here, as long as the weights of multiple first request information sets in the first part are generated according to the Mth weight. The number of weights generated according to the Mth weight is the same as the number of first request information sets in the first part. The weights generated by the Mth weight may be randomly assigned to the first request information sets in the first part, and each first request information set in the first part can match a new weight, update the existing first weight, and thus realize the update of the first weight.

[0123] For example, according to the third weight of the third request information set in the third part, the first weights corresponding to the 10 first request information sets included in the first part are updated. Specifically, 10 first weights may be generated according to the third weight, and according to the 10 newly generated first weights, the first weights already possessed by the 10 first request information sets in the first part are updated.

[0124] For step S700, the first target information matching the request information in the first request information set is re-determined from the information to be recommended according to the updated first weight. Since the first weight is re-updated, the target information that better matches the request information in the first request information set can be re-determined from the information to be recommended according to the updated weight, that is, the first target information. The target information matching the request information in the Nth request information set is recorded as the Nth target information.

[0125] After cyclically updating each weight, each weight can be automatically updated to achieve full automatic update, eliminating the process of manually updating each weight offline, thereby reducing the update difficulty caused by manually updating each weight offline, which in turn leads to untimely updates of each weight, further affecting the occurrence of inaccurate recommended target information. Through this method, each weight can be automatically and timely updated, and more matching target information can be recommended to the request information in the corresponding request information set according to the updated weight, further improving the accuracy of the recommended target information and improving the user experience.

[0126] In another embodiment, step S600, after updating the Mth weight, updating the first weights corresponding to the plurality of first request information sets included in the first part according to the updated Mth weight, includes:

[0127] According to the updated Mth weight, target weights having the same number as the first request information set included in the first part are determined, and then the first weight is updated according to the target weight.

[0128] Specifically, the new first weight, i.e., the target weight, can be determined by Gaussian distribution according to the updated Mth weight. The specific process can be to randomly generate new values ​​through Gaussian distribution and determine them through mean and variance. The current Mth weight is selected as the mean through Gaussian distribution, and the variance can be set according to actual needs. The target weight can be determined through Gaussian distribution according to the mean and variance. The smaller the variance, the closer the value of the target weight randomly determined by Gaussian distribution is to the mean, that is, it fluctuates around the mean. The larger the variance, the farther the target weight randomly determined by Gaussian distribution is from the mean, that is, it is distributed farther away from the mean. In combination with actual business needs, you can choose to randomly determine the target weight near the mean, and the variance is small at this time. You can also determine the target weight from more locations as needed, and the variance is larger at this time.

[0129] For example, 10 target weights are determined by the above method, and the first weights already possessed by the 10 first request information sets in the first part are updated by the 10 target weights. The 10 target weights and the 10 existing first weights can be randomly assigned, and each target weight updates an existing first weight.

[0130] In another embodiment, after updating the first weights corresponding to the plurality of first request information sets included in the first part, the method further includes:

[0131] According to the updated first weight, the second weight is updated to the Mth weight; according to the updated Nth weight, the N+1th weight is updated.

[0132] After updating the first weight, update the existing second weight according to the updated weight, update the existing third weight according to the updated second weight, and so on, update the existing Mth weight according to the updated M-1th weight, and so on, to achieve automatic update of the weights of each request information set in each part.

[0133] In another embodiment, in step S100, the request information of multiple different users is divided into M parts, including:

[0134] According to the receiving time of the request information, the request information received in the same time period is divided into the same part; and / or, the sorting value is obtained according to the priority of the user sending the request information and the sorting factor corresponding to the receiving time of the request information; the request information with sorting values ​​in the same interval is divided into the same part.

[0135] A method for dividing the request information according to the time when the request information is received. When dividing the request information of multiple different users, the request information can be divided with reference to the time when the request information is received, and the request information received in the same time period can be divided into the same part, so that the recommended target information can be determined according to the time information used by the user.

[0136] For example, the request information received from 8:00 a.m. to 10:00 a.m. is divided into the same part, such as the first part. The request information received from 12:00 p.m. to 1:00 p.m. is divided into the same part, such as the second part. The request information received from 8:00 p.m. to 11:00 p.m. is divided into the same part, such as the third part. The total number of request information included in each request information set of the first part is less than the total number of request information included in each request information set included in the second part; the total number of request information included in each request information set included in the second part is less than the total number of request information included in each request information set included in the third part. Moreover, the number of request information included in each request information set included in the first part is less than the number of request information included in each request information set included in the second part, and the number of request information included in each request information set included in the second part is less than the number of request information included in each request information set included in the third part.

[0137] This can facilitate updating the weight of a set including a majority of request information by using the weight of the set including a minority of request information. When an abnormality occurs in target information recommended to the request information in the set including a minority of request information based on the weight used for the set including a minority of request information, the impact on the weight of the set including a majority of request information can be reduced, thereby reducing the impact on the accuracy of determining the target information recommended to the request information in the set including a majority of request information.

[0138] Another method is to divide according to the receiving time and the user priority. The priority of the user who sends the request information can be determined according to the preset priority, for example, the priority of the user who opens the membership is higher than the priority of the ordinary user.

[0139] The sorting factor can be set as needed. The priority of the user sending the request information and the time of receiving the request information can correspond to the sorting factors respectively. Different sorting factors correspond to different values. The final sorting value can be determined based on the sorting factors of the two. The division of the request information is determined based on the sorting value, and the request information with sorting values ​​in the same interval can be divided into the same part.

[0140] Exemplarily, according to the order of receiving time of the request information, the earlier the receiving time is, the greater the value of the first calculation factor is; according to the user priority, the higher the priority is, the greater the value of the second calculation factor is.

[0141] The receiving time corresponds to a first weight factor, and the priority corresponds to a second priority;

[0142] The ranking factor may be obtained based on the first calculation factor, the first weight factor, the second priority and the second weighted average.

[0143] As another example, the number of request information of each priority user is determined according to the priority level, and the request information of the same priority user is divided into a request information set. If the number of request information contained in the request information set is not enough, part or all of the request information of the next priority user is divided into the request information set. When dividing part of the request information of the next priority into the request information set, according to the reception time of the request information of the next priority user, the part of the request information with the earlier reception time is divided into the current request information set.

[0144] In the embodiments of the present disclosure, the request information sets responded to at different time points will comprehensively consider the user priority and receiving time of the request information, so as to achieve average response delay control while taking into account the different requirements of different priority users for response delay.

[0145] The method uses the priority of the user sending the request information and the receiving time of the request information as reference factors to determine how to divide the request information.

[0146] In another embodiment, reference Figure 5 , is a schematic diagram of a structure of a device for recommending target information, the device comprising:

[0147] A division module 1 is used to divide the request information of multiple different users into M parts, the Nth part includes multiple Nth request information sets, the N+1th part includes at least one N+1th request information set, and the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set; wherein N is greater than or equal to 1 and less than or equal to M-1, and N and M are both positive integers;

[0148] An Nth target information determination module 2, configured to determine, from the information to be recommended, Nth target information matching the request information in the Nth request information set according to an Nth weight of at least one dimension information of the information to be recommended;

[0149] Feedback information acquisition module 3, used to obtain feedback information after the Nth target information is recommended;

[0150] An updating module 4, configured to update the N+1th weight of at least one dimension of the information to be recommended according to the feedback information;

[0151] The N+1th target information determining module 5 is used to determine, from the information to be recommended, the N+1th target information matching the request information in the N+1th request information set based on the updated N+1th weight.

[0152] In another embodiment, the Mth part includes: a set of the Mth request information;

[0153] The device also includes:

[0154] A first weight updating module, configured to update, after updating the Mth weight, first weights corresponding to the plurality of first request information sets included in the first part according to the updated Mth weight;

[0155] The first target information redetermining module is used to redetermine, from the information to be recommended, first target information that matches the request information in the first request information set according to the updated first weight.

[0156] In another embodiment, the first weight updating module is specifically used to: determine, through a cross entropy algorithm, the same number of target weights as the first request information set included in the first part based on the updated Mth weight; and update the first weight based on the target weight.

[0157] In another embodiment, the device also includes a second updating module, which is used to update the second weight to the Mth weight according to the updated first weight after updating the first weights corresponding to the multiple first request information sets included in the first part; wherein the N+1th weight is updated according to the updated Nth weight.

[0158] In another embodiment, the plurality of request information in the Nth request information set have the same Nth weight;

[0159] Feedback information acquisition module 3, including:

[0160] The first determination module is used to determine the feedback dimension information of the Nth target information and the parameters of each feedback dimension information.

[0161] The second determination module is used to determine the feedback value according to the feedback dimension information and the parameters of each feedback dimension information.

[0162] In another embodiment, the update module 4 includes:

[0163] A feedback value determination module, used to determine the feedback values ​​respectively corresponding to the request information in the plurality of Nth request information sets;

[0164] An Nth target weight determination module, configured to determine an Nth weight corresponding to the request information in the Nth request information set having the largest feedback value as the Nth target weight;

[0165] The N+1th weight updating module is used to update the N+1th weight according to the Nth target weight.

[0166] In another embodiment, the feedback dimension information includes at least one of the following:

[0167] The number of downloads, collections, likes, number of users following the author of the recommended information, average browsing time, number of comments, and number of reposts.

[0168] In another embodiment, the partitioning module 1 includes:

[0169] A first division module, configured to divide the request information received at the same time period into the same part according to the time at which the request information is received;

[0170] and / or,

[0171] The second division module is used to obtain a ranking value according to the priority of the user sending the request information and the ranking factor corresponding to the receiving time of the request information; and divide the request information with ranking values ​​in the same interval into the same part.

[0172] In another embodiment, the information to be recommended includes at least one of the following: video; public account; article; advertisement;

[0173] The dimensional information includes at least one of the following: download volume, collection volume, like volume, number of users following the author of the recommended information, average browsing time, number of comments, and number of reposts.

[0174] In another embodiment, an electronic device is provided, including:

[0175] A processor and a memory for storing executable instructions capable of running on the processor, wherein:

[0176] When the processor is used to run the executable instructions, the executable instructions execute the method described in any one of the above embodiments.

[0177] In another embodiment, a non-transitory computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, the method described in any of the above embodiments is implemented.

[0178] In another embodiment, another method for recommending target information is provided.

[0179] In today's information age, the amount of data is increasing day by day. In the face of a variety of choices in the massive data, it is often difficult to determine the information you need. Therefore, in order to make reasonable choices in the massive data and determine the information you need, the recommendation system was born. Taking video recommendation as an example, the recommendation system uses the user's historical behavior data, user portrait data, and statistical data characteristics on the user side or the video side, and uses the corresponding recommendation algorithm to determine the information that the user may be interested in. With the update of the recommendation system, the recommendation accuracy is getting higher and higher, but the corresponding problem is the need for a more complex recommendation system and updated experience.

[0180] Especially in the current personalized recommendation scenario of information flow, recommendation algorithms are changing with each passing day, and the popular multi-objective recommendation model is now widely used. There is less displayed feedback in the recommendation system because most user feedback is not a direct rating. Most recommendation systems recommend based on implicit feedback, such as user attention, sharing, likes, viewing time, and complete playback. There are some cognitive biases when evaluating user satisfaction, mainly including target bias, item bias, user bias, etc. In order to solve the target bias, item bias, user bias, etc., a multi-model score fusion method is adopted. Each optimization target is predicted by an independent model to obtain a score, and finally these scores are fused in a weighted manner. Each model has a score, and a final score is designed by the weight of the model to complete multi-objective optimization.

[0181] Generally, there are various methods used for multi-objective parameter adjustment technology, including manual parameter adjustment and automatic parameter adjustment. However, whether it is manual parameter adjustment or automatic parameter adjustment, these parameter adjustment methods are offline adjustments and then updated online, and it is impossible to completely break away from human intervention to achieve the iteration of online automatic parameter adjustment. In addition, although manual parameter adjustment is lightweight and has good interpretability, it relies on the experience of engineers, and automatic parameter adjustment requires more experience for the analysis of the relationship between multiple targets. Especially when the parameters are expanded to 10, 20, or even more, offline adjustment will become more and more difficult. So if this set of parameters can get rewards based on their own parameter changes and then interact with the environment, obtain new information, and then change themselves to get rewards. In this way, these parameters can change themselves by getting rewards, move in a better direction, and get rid of human intervention.

[0182] In the recommendation system, request information includes traffic, which refers to the customer's access request. Taking the video recommendation system as an example, for each user's request information, the system will return the video to the user for viewing based on the information in the request and offline data through multiple layers and models. Traffic division means dividing the customer's access request information into M parts, each of which has multiple sets of request information. For example, each part includes several buckets, and the traffic of each bucket can have its own system black box, which can have different model parameters. The final choice can be determined based on the feedback in different black boxes. This solution chooses the form of feedback from small traffic to large traffic selection, mainly for two reasons. On the one hand, the small traffic experiment will only affect a small number of users, and once a problem occurs, it will not cause a large-scale problem. On the other hand, each of our traffic buckets is randomly sampled, assuming that they meet the normal distribution, and the overall effect is fed back with the effect of small traffic.

[0183] refer to Figure 6 , which is a schematic diagram of another method.

[0184] Step 1: Divide 100% of the traffic into several small traffics, and then determine the selection of the large traffic according to the feedback of the small traffic, forming a feedback parameter update method. For example, divide 100% of the traffic into three parts, including the third part, the second part and the first part. The third part includes a request information set including 30% of the request information (also called 30% traffic), the second part includes two request information sets including 20% ​​of the request information respectively (also called 20% traffic), and the first part includes 10 request information sets including 3% of the request information respectively (also called 3% traffic). The total amount of request information included in all request information sets is 100%. The feedback effect of the target information determined by the parameters in several 3% request information sets can be used to determine how to determine and update the parameters in the 20% request information set, and then the parameters of the 30% request information set can be determined by the parameters of the 20% request information set.

[0185] Step 2: Divide the 10 request information sets including 3% into two groups, each group has 5 request information sets including 3%. Each 3% request information set in each group will have its own set of parameters (a, b, c, etc.), such as an*likes+bn*completed+cn*shares+etc. (n=1,2,3,4,5). Based on the parameters, the target information recommended to each request information set including 3% can be determined. Then, based on the feedback information of the target information, for example, the feedback information is determined based on the reward equation reward=p*content duration+t*average playback duration+e*average number of playbacks+etc., and the content duration, average playback duration and average number of playbacks are used as feedback dimension information. Obtain the parameters an, bn, cn, etc. of a set of request information sets including 3% with the best feedback effect.

[0186] Step 3: Update the parameters of the request information set that includes 20% according to the parameters of the set of request information that has the best feedback effect, which includes 3%. Since the request information set that includes 3% is divided into two large groups, two sets of parameters of the set of request information that has the best feedback effect, which includes 3%, will be obtained, and then the parameters of the two sets of request information sets that include 20% will be updated respectively.

[0187] Step 4: Based on the feedback information of the two groups of 20% request information sets, select the parameters of the group of 20% request information sets with the best feedback effect, and update the parameters of the 30% request information sets.

[0188] Step 5: Based on the feedback information of the target information including 30% of the request information set, for example, the reward equation reward is obtained. If it is better than the feedback effect of the previous group, the new parameters are used. Then, through the cross entropy algorithm, using the current parameters, 10 sets of parameters are selected and updated to the parameters of the 10 request information sets including 3% of the request information.

[0189] Step 6: Loop through the above steps, continuously iterate the parameters of each request information set, and complete the continuous iteration of the parameters.

[0190] It should be noted that the “first” and “second” in the embodiments of the present disclosure are only for the convenience of description and distinction and have no other specific meanings.

[0191] Figure 7 1 is a block diagram of a terminal device according to an exemplary embodiment. For example, the terminal device 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.

[0192] Reference Figure 7 The terminal device may include one or more of the following components: a processing component 802 , a memory 804 , a power component 806 , a multimedia component 808 , an audio component 810 , an input / output (I / O) interface 812 , a sensor component 814 , and a communication component 816 .

[0193] The processing component 802 generally controls the overall operation of the terminal device, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above method. In addition, the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0194] The memory 804 is configured to store various types of data to support operations on the terminal device. Examples of such data include instructions for any application or method operating on the terminal device, contact data, phone book data, messages, pictures, videos, etc. The memory 804 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.

[0195] The power component 806 provides power to various components of the terminal device. The power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the terminal device.

[0196] The multimedia component 808 includes a screen that provides an output interface between the terminal device 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 808 includes a front camera and / or a rear camera. When the terminal device 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.

[0197] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC), and when the terminal device is in an operation 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 804 or sent via the communication component 816. In some embodiments, the audio component 810 also includes a speaker for outputting audio signals.

[0198] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: home button, volume button, start button, and lock button.

[0199] The sensor assembly 814 includes one or more sensors for providing various aspects of status assessment for the terminal device. For example, the sensor assembly 814 can detect the open / closed state of the terminal device, the relative positioning of components, such as the display and keypad of the terminal device, and the sensor assembly 814 can also detect the position change of the terminal device or a component of the terminal device, the presence or absence of user contact with the terminal device, the orientation or acceleration / deceleration of the terminal device, and the temperature change of the terminal device. The sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 814 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 814 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0200] The communication component 816 is configured to facilitate wired or wireless communication between the terminal device and other devices. The terminal device 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 816 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 816 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.

[0201] In an exemplary embodiment, the terminal device 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 method.

[0202] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0203] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for recommending target information, characterized in that: include: Divide the request information of multiple different users into M parts, the Nth part includes multiple Nth request information sets, the N+1th part includes at least one N+1th request information set, and the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set; wherein N is greater than or equal to 1 and less than or equal to M-1, and both N and M are positive integers; determining, from the information to be recommended, Nth target information matching the request information in the Nth request information set according to an Nth weight of at least one dimension information of the information to be recommended; Obtain feedback information after the Nth target information is recommended; According to the feedback information, updating the N+1th weight of at least one dimension information of the information to be recommended; Based on the updated N+1th weight, determine the N+1th target information matching the request information in the N+1th request information set from the information to be recommended.

2. The method according to claim 1, characterized in that The Mth part includes: an Mth request information set; The method further comprises: After updating the Mth weight, updating the first weights corresponding to the plurality of first request information sets included in the first part according to the updated Mth weight; According to the updated first weight, first target information matching the request information in the first request information set is re-determined from the information to be recommended.

3. The method according to claim 2, characterized in that After updating the Mth weight, updating the first weights corresponding to the plurality of first request information sets included in the first part according to the updated Mth weight, comprises: Determine, according to the updated Mth weight, the same number of target weights as the first request information set included in the first part; The first weight is updated according to the target weight.

4. The method according to claim 2, characterized in that: After updating the first weights corresponding to the plurality of first request information sets included in the first part, the method further includes: According to the updated first weight, the second weight is updated to the Mth weight; wherein, according to the updated Nth weight, the N+1th weight is updated.

5. The method according to claim 1, characterized in that The plurality of request information in the Nth request information set have the same Nth weight; The obtaining feedback information after the Nth target information is recommended includes: Determining feedback dimension information of the Nth target information and parameters of each feedback dimension information; A feedback value is determined according to the feedback dimension information and parameters of each of the feedback dimension information.

6. The method according to claim 5, characterized in that The updating, according to the feedback information, the N+1th weight of at least one dimension of the information to be recommended includes: Determine the feedback values ​​respectively corresponding to the request information in the plurality of request information sets; Determine the Nth weight corresponding to the request information in the Nth request information set having the largest feedback value as the Nth target weight; The N+1th weight is updated according to the Nth target weight.

7. The method according to claim 5, characterized in that The feedback dimension information includes at least one of the following: The number of downloads, collections, likes, number of users following the author of the recommended information, average browsing time, number of comments, and number of reposts.

8. The method according to claim 1, characterized in that The dividing the request information of multiple different users into M parts includes: According to the receiving time of the request information, the request information received at the same time period is divided into the same part; and / or, A ranking value is obtained according to the priority of the user sending the request information and the ranking factor corresponding to the receiving time of the request information; and the request information with ranking values ​​in the same interval is divided into the same part.

9. The method according to any one of claims 1 to 8, characterized in that: The information to be recommended includes at least one of the following: video; Official account; article; advertise; The dimension information includes at least one of the following: The number of downloads, collections, likes, number of users following the author of the recommended information, average browsing time, number of comments, and number of reposts.

10. A device for recommending target information, characterized in that: include: A division module, used to divide the request information of multiple different users into M parts, the Nth part includes multiple Nth request information sets, the N+1th part includes at least one N+1th request information set, and the number of request information included in the Nth request information set is less than the number of request information included in the N+1th request information set; wherein N is greater than or equal to 1 and less than or equal to M-1, and N and M are both positive integers; An Nth target information determination module, configured to determine, from the information to be recommended, Nth target information matching the request information in the Nth request information set according to an Nth weight of at least one dimension information of the information to be recommended; A feedback information acquisition module, used to acquire feedback information after the Nth target information is recommended; An updating module, configured to update the N+1th weight of at least one dimension of the information to be recommended according to the feedback information; The N+1th target information determination module is used to determine, from the information to be recommended, the N+1th target information that matches the request information in the N+1th request information set based on the updated N+1th weight.

11. An electronic device, characterized in that: include: A processor and a memory for storing executable instructions capable of running on the processor, wherein: When the processor is used to run the executable instructions, the executable instructions execute the method described in any one of claims 1 to 9.

12. A non-transitory computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the method described in any one of claims 1 to 9.

Images