CN109408703B - Information recommendation method and system, device, electronic equipment and storage medium thereof - Google Patents
Information recommendation method and system, device, electronic equipment and storage medium thereof Download PDFInfo
- Publication number
- CN109408703B CN109408703B CN201811020544.9A CN201811020544A CN109408703B CN 109408703 B CN109408703 B CN 109408703B CN 201811020544 A CN201811020544 A CN 201811020544A CN 109408703 B CN109408703 B CN 109408703B
- Authority
- CN
- China
- Prior art keywords
- information
- recall
- user
- data
- services
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The utility model discloses an information recommendation method, system, device, electronic equipment and computer readable storage medium, the scheme includes: acquiring corresponding user data according to information recommendation triggered by a user; triggering to recall information through a plurality of information recall services which are acquired and requested to be deployed by user data in parallel; the information recall service recommends and executes information recall for the user information from the configured positive data set according to the user data to obtain a return result of the information recall service, wherein the positive data set is formed by segmenting all positive data; and summarizing the returned results of all the information recall services to form an information recommendation result. The scheme reduces the time for completing the information recall task, improves the information recommendation efficiency and supports the great increase of the data volume.
Description
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to an information recommendation method, an information recommendation system, an information recommendation apparatus, an electronic device, and a computer-readable storage medium.
Background
With the rapid development of network technologies, a great deal of information is generated on the internet every day, and therefore, how to quickly find out network information that a user wants or is suitable for the user from massive data becomes a technical problem that technical personnel urgently need to solve.
At present, the intelligent recommendation system mainly pushes information matched with basic attributes and historical behaviors of a user according to some basic attributes (including age, gender and preference) and the historical behaviors of the user. For example, a favorite article is obtained for the user according to the reading habit and the interest of the user. And recommending articles which are popular with all the users of the same kind to the users who do not read the articles. Alternatively, the same type of article is recommended to unread users.
Therefore, the recommendation system depends on a great deal of data, including the historical behaviors and basic attributes of all users and the detailed data of various articles, and as the number of the users increases and time is accumulated, the data is increased, so that the data processing takes longer and the response of the recommendation system is slow.
Disclosure of Invention
In order to solve the problems that a recommendation system in the related art is long in time consumption and slow in response, the information recommendation method is provided.
In one aspect, the present invention provides an information recommendation method, including:
acquiring corresponding user data according to information recommendation triggered by a user;
triggering a plurality of information recall services deployed according to the acquired user data parallel request to recall information;
the information recall service recommends and executes information recall for the information of the user from the configured front-ranking data set according to the user data to obtain a return result of the information recall service, wherein the front-ranking data set is formed by segmenting all front-ranking data;
and summarizing the returned results of all the information recall services to form an information recommendation result.
In another aspect, the present invention further provides an information recommendation system, including: the system comprises a proxy server, a scheduling server, a fusion server and a plurality of recall servers which are connected in sequence;
the proxy server is used for receiving a request for information recommendation to a user;
the scheduling server is used for acquiring corresponding user data according to the user identification indicated by the request;
the fusion server is used for requesting a plurality of recall servers to recall information in parallel through the user data;
the recall server is used for operating information recall service, the information recall service recommends and executes information recall for the information of the user from the configured front-row data set according to the user data to obtain a return result of the information recall service, and the front-row data set is formed by segmenting all the front-row data;
the scheduling server is also used for summarizing the return results of all the information recall services to form an information recommendation result.
In addition, the present invention also provides an information recommendation apparatus, comprising:
the data acquisition module is used for acquiring corresponding user data according to information recommendation triggered by the user;
the parallel request module is used for triggering information recall through a plurality of acquired information recall modules which are deployed according to the user data parallel request;
the information recall module is used for recommending and executing information recall for the information of the user from the configured main data set according to the user data to obtain a return result of the information recall module, wherein the main data set is formed by segmenting all main data;
and the result summarizing module is used for summarizing the returned results of all the information recalling modules to form an information recommendation result.
Further, the present invention also provides an electronic device, including:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to execute the information recommendation method.
The invention also provides a computer-readable storage medium, which stores a computer program, wherein the computer program can be executed by a processor to implement the information recommendation method.
The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects:
according to the technical scheme provided by the invention, all the front-ranking data are segmented to form a plurality of front-ranking data sets, so that a plurality of information recall services can execute information recall on the deployed front-ranking data sets in parallel, and the returned results of all the information recall services are summarized to obtain the information recommendation result. Therefore, the data volume loaded by each information recall service can be reduced, the time consumed by processing is reduced, a plurality of information recall services carry out information recall in parallel, the time for completing the information recall task is reduced, the information recommendation efficiency is improved, and the data volume is greatly increased.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a schematic illustration of an implementation environment according to the present disclosure;
FIG. 2 is a block diagram illustrating a server in accordance with an exemplary embodiment;
FIG. 3 is a flow diagram illustrating a method of information recommendation in accordance with an exemplary embodiment;
FIG. 4 is a diagram of a prior art architecture for decomposing a recall service by function;
FIG. 5 is a flowchart showing details of step 350 in the corresponding embodiment of FIG. 3;
FIG. 6 is a detailed flowchart of step 352 in a corresponding embodiment of FIG. 5;
FIG. 7 is a flowchart showing details of step 370 in the corresponding embodiment of FIG. 3;
FIG. 8 is an architectural diagram illustrating a server cluster in accordance with an exemplary embodiment;
FIG. 9 is a block diagram illustrating a stand-alone server in accordance with an exemplary embodiment;
FIG. 10 is an architectural diagram of an information recommendation system shown in accordance with an exemplary embodiment;
FIG. 11 is a block diagram illustrating an information recommendation device in accordance with an exemplary embodiment;
FIG. 12 is a detailed block diagram of an information recall module in a corresponding embodiment of FIG. 11;
FIG. 13 is a block diagram of the details of the result acquisition unit in the corresponding embodiment of FIG. 12;
FIG. 14 is a block diagram illustrating details of a result summarization module in the corresponding embodiment of FIG. 11.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like 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 invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the invention, as detailed in the appended claims.
FIG. 1 is a schematic illustration of an implementation environment according to the present disclosure. The implementation environment includes a server cluster of servers 110. The plurality of servers 110 are connected by a wired or wireless network. The server cluster can adopt the information recommendation method provided by the invention to generate the information recommendation result.
Further, the implementation environment may further include a mobile terminal 130, and the mobile terminal 130 may send a request for information recommendation to a user to the server cluster, and then the server cluster responds to the request, and returns an information recommendation result to the mobile terminal 130 by using the information recommendation method provided by the present invention.
Referring to fig. 2, fig. 2 is a schematic diagram of a server structure according to an embodiment of the present invention. The server 200 may vary significantly depending on configuration or performance, and may include one or more Central Processing Units (CPUs) 222 (e.g., one or more processors) and memory 232, one or more storage media 230 (e.g., one or more mass storage devices) storing applications 242 or data 244. Memory 232 and storage medium 230 may be, among other things, transient or persistent storage. The program stored in the storage medium 230 may include one or more modules (not shown), each of which may include a series of instruction operations for the server 200. Still further, the central processor 222 may be configured to communicate with the storage medium 230 to execute a series of instruction operations in the storage medium 230 on the server 200. The Server 200 may also include one or more power supplies 226, one or more wired or wireless network interfaces 250, one or more input-output interfaces 258, and/or one or more operating systems 241, such as a Windows Server TM ，Mac OS XTM ，UnixTM,Linux TM ，FreeBSD TM And so on. The steps performed by the server cluster described in the embodiments of fig. 3, 5-7 below may be based on the server architecture shown in fig. 2.
It will be understood by those skilled in the art that all or part of the steps for implementing the following embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
It should be noted that, in the embodiments of the present application, the data related to the user is referred to, when the embodiments of the present application are applied to specific products or technologies, user permission or consent needs to be obtained, and the collection, use and processing of the related data need to comply with the relevant laws and regulations and standards of the relevant countries and regions.
Fig. 3 is a flow chart illustrating an information recommendation method according to an example embodiment. The application scope and the execution subject of the information recommendation method may be, for example, a server cluster in the implementation environment shown in fig. 1. As shown in fig. 3, the method may be performed by a cluster of servers and may include the following steps.
In step 310, according to the information recommendation triggered by the user, acquiring the corresponding user data;
wherein, the information recommendation is directed to the user to recommend information which may be interested in the user. The information may be in the form of, for example, images, text, video (including live video), and the like. The user data is used to characterize the basic characteristics of the user. The user data includes user portrait information and user behavioral data. The user portrait information is a tagged user model abstracted according to the information of the user social attribute, the living habits, the consumption behaviors and the like, namely, the user portrait information can be used for representing the preference of the user. And the user behavior data records the information that the user has clicked, and can be history browsing records.
In an embodiment, the information recommendation triggered by the user may be a request sent by the mobile terminal to the server cluster in response to the trigger, for indicating to recommend information to the user to which the mobile terminal belongs. And the server cluster acquires the user data stored corresponding to the user identification according to the user identification carried by the request. In another embodiment, the information recommendation triggered by the user may be a triggering instruction for recommending information to the user, which is detected by the server cluster. And the server cluster acquires the user data stored corresponding to the user identification according to the user identification carried by the trigger instruction.
In step 330, information recall is triggered by a plurality of information recall services which are requested to be deployed by the acquired user data in parallel;
the information recall service refers to a program which is deployed on a server and used for executing information recall. Each server may deploy one or more information recall services. And a plurality of servers can realize the decentralized deployment of a large amount of information recall services. Each server can recall information by running the information recall service deployed by the server. It should be explained that the information recall refers to obtaining information, which is interesting to the user and has a high possibility of clicking, from the information collection.
In one embodiment, a server cluster may include a master server, a plurality of slave servers. And the master server acquires corresponding user data according to the information recommendation triggered by the user, and then sends recall requests carrying the user data to the plurality of slave servers in parallel. And sending the recall request to a plurality of slave servers, triggering the plurality of slave servers to run self-deployed information recall services, and acquiring information which is interesting to the user and has high click possibility from the self-configured information set.
In step 350, the information recall service recommends and executes information recall for the user information from the configured positive data set according to the user data to obtain a return result of the information recall service, wherein the positive data set is formed by segmenting all positive data;
the main data refers to detail data of information (such as articles), including titles, time, topics, and the like. All the positive rank data, namely the detail data of all the information, can be divided into a plurality of positive rank data sets. The positive data set may be considered as an information set composed of a plurality of pieces of information, and thus, all the positive data sets are divided into a plurality of positive data sets, and all the information sets may be considered as a plurality of information sets, each of which includes a plurality of pieces of information.
The returned result of the information recall service refers to information which is acquired by the information recall service from the positive data set and is possibly interested in the user. Depending on the actual situation, the returned result may contain a plurality of pieces of information.
It should be noted that the information recall service may determine information that may be of interest to the user based on the user data. In one embodiment, the plurality of positive data sets obtained by segmenting all the positive data are distributed and deployed on the plurality of slave servers, so that each slave server can obtain information which is possibly interested by the user from the configured positive data sets according to the user data by running the information recall service.
In step 370, the returned results of all the information recall services are summarized to form an information recommendation result.
The summarizing of the returned results of all the information recall services may be simply combining all the returned results, or summarizing all the returned results according to a configured policy and a specified policy. The information recommendation result refers to information finally pushed to the user.
In an embodiment, each slave server of the server cluster sends information (i.e., a return result) that is acquired by the running information recall service and is likely to be interested by the user to the master server, and the master server of the server cluster may further summarize the return result sent by each slave server, that is, summarize all information that is returned by the slave servers and is likely to be interested by the user, so as to obtain information that is finally pushed to the user. Furthermore, the server cluster can send the determined information which needs to be pushed to the user finally to the mobile terminal to which the user belongs, and the mobile terminal displays the pushed information at the specified position.
It should be noted that, the present invention divides all the main data into a plurality of main data sets, and executes information recall from the plurality of main data sets in parallel through a plurality of deployed information recall services. Each information recall service performs the same recall logic. Therefore, even if the number of the front row data is increased, the level expansion of the information recall service can be performed, and the newly added front row data set is processed in parallel by further adding the information recall service, so that the response time of information recommendation can not be influenced.
In the existing scheme, the recall service is mainly decomposed according to functional modules and split into services with different functions, and each service executes different recall logics. As shown in FIG. 4, the recall service is broken down into interest recalls, collaborative recalls, extended recalls, and other recalls by functional module. The interest recall means that articles which are liked by the user are recalled for the user according to the reading habit interest of the user. The collaborative recall refers to recommending articles which are popular with people in the same type of users to users who do not read the articles. Or recommending to unread users with articles of the same type. The expanding recall refers to recalling information except the interest of the user to expand the interest of the user so as to more comprehensively discover the interest of the user on various information. The interest recall, collaborative recall and extended recall services respectively execute different recall logics, so the development cost is high. Each service recalls information from all the positive data, and the response time is slower and slower when the positive data is more and more. The probability of inconsistency of the information recalled by each service is high, and the information recommendation is easy to make mistakes.
According to the technical scheme provided by the above exemplary embodiment of the invention, all the front-ranking data are segmented to form a plurality of front-ranking data sets, then a plurality of information recall services can execute information recall on the deployed front-ranking data sets in parallel, and the returned results of all the information recall services are summarized to obtain the information recommendation result. Therefore, the data volume loaded by each information recall service can be reduced, the processing time is reduced, a plurality of information recall services are used for recalling information in parallel, the time for completing the information recall task is reduced, the information recommendation efficiency is improved, and the data volume is greatly increased.
In an exemplary embodiment, the step 330 specifically includes:
and sending the acquired user data to a plurality of deployed information recall services in parallel, wherein the sending of the user data to the information recall services triggers the information recall services to execute information recall for users.
Specifically, the server cluster includes a master server and a plurality of slave servers, where the master server acquires user data and sends requests carrying the user data to the plurality of slave servers in parallel. And each slave server is deployed with an information recall service, the sending of the user data to the slave servers triggers the slave servers to run the information recall service, and the information related to the user data is acquired from the self-deployed forward data set.
In an exemplary embodiment, the information recommendation of the invention deploys a plurality of vertical services adapted to a plurality of recommendation information types, and a plurality of information recall services are deployed for information recall of corresponding recommendation information types at the downstream of the vertical services; the steps are as follows: sending the acquired user data to a plurality of deployed information recall services in parallel, comprising: by scheduling several vertical services, user data is sent in parallel to multiple information recall services deployed downstream of each vertical service.
The recommendation information type refers to the data type of the information, and can be divided into: graphics, video clips, live video, and the like. The vertical service is a computer program which sends user data to a plurality of information recall services and fuses return results of the plurality of information recall services according to a specified adjustment strategy. For each data type of information, there is a corresponding vertical service. For example with teletext services for teletext information. Video services for video clips, live services for live video, and so on. The plurality of vertical services include a text-to-text fusion service, a video service, a live broadcast service, and the like.
A plurality of information recall services are deployed downstream of each vertical service, for example, a plurality of information recall services are deployed downstream of the image-text fusion service and are used for performing information recall on image-text information. A plurality of information recall services are deployed downstream of the video service for performing information recall on the video segments. A plurality of information recall services are deployed downstream of the live broadcast service for performing information recall on live video.
In particular, each vertical service may be deployed on one or more servers. For differentiation, the server in which the vertical service is deployed is referred to as a scheduling server, and the scheduling server may be regarded as belonging to the master server in the above exemplary embodiment, and the server in which the information recall service is deployed is referred to as a recall server, i.e., a slave server in the above exemplary embodiment. The invocation server sends requests containing user data to a plurality of information recall services downstream of each vertical by scheduling the plurality of vertical services. For example, the scheduling server schedules the teletext service and sends requests carrying user data to a plurality of information recall services (for recall of teletext information) downstream of the teletext service. The scheduling server schedules the video service to send a request carrying user data to a plurality of information recall services (for recall of video segments) downstream of the video service. Similarly, as the types of the information data increase, the scheduling server may also schedule other services to send requests carrying user data to a plurality of information recall services (for recalling other information) downstream.
In an exemplary embodiment, as shown in fig. 5, the step 350 specifically includes:
in step 351, the information recall service screens out target information with information characteristics matched with the user data from the configured forward data set according to the user data;
wherein, each information recall service is correspondingly configured with a positive data set. The information recall service within the same server loads the same positive data set. The positive rank data set is obtained by segmenting all positive rank data according to the number of information recall services. The forward data refers to the detail data of each piece of information, and thus, the forward data set is a set of several pieces of information and contains the detail data of several pieces of information.
It should be noted that the information features are used to represent characteristics of different information, including information belonging fields, crowd-oriented characteristics, propagation areas, and the like. And the user data is used to indicate the user's hobbies and historical behavior. The matching of the information characteristic with the user data means that the information characteristic meets the requirements indicated by the user data. For example, the information characteristic indicates that the information belongs to the field of sports, and the user data indicates that the user's interests include sports, and the information characteristic and the user data may be considered to match. For example, an information feature indicating entertainment news belonging to a certain story, and user data indicating that the user's idol belongs to a certain story, the information feature may be considered a match with the user data. The target information refers to information which is screened from the forward data set and has information characteristics matched with the user data.
Specifically, each recall server runs the information recall service deployed by itself, and each information recall service can screen out information with information characteristics matched with user data from the corresponding configured forward data set according to the received user data to serve as target information. The target information may be one or more pieces of information.
In step 352, a return result of the information recall service is obtained according to the target information.
Specifically, the information recall service may perform further filtering processing on the target information according to the target information obtained from the forward data set, to obtain information returned to the upstream vertical service, that is, a returned result.
In an exemplary embodiment, as shown in fig. 6, the step 352 specifically includes:
in step 3521, the information recall service calculates the click rate of the user to which the user data belongs to the target information by using the constructed prediction model;
the prediction model can be obtained by training a large amount of sample data in advance. And the information recall service combines the information characteristics of the target information and the user data aiming at each piece of target information, inputs the constructed prediction model and outputs the click rate of the user to which the user data belongs to the target information. The click rate is used for representing the possibility that the user to which the user data belongs clicks the target information, namely the possibility that the user can read the target information.
In step 3522, the target information is sorted according to the click rate, and the return result of the information recall service is determined according to the sorted result.
Specifically, the information recall service sorts all target information according to the click rate of each piece of target information by the user and the size of the click rate. Specifically, all the target information may be sorted from large to small according to the click rate, and the n item label information sorted in the top may be used as the return result of the information recall service. n can be set according to actual needs. That is, the top n items of the mark information with the largest click rate are used as the return result of the information recall service.
In one embodiment, before step 3521, the information recommendation method provided by the present invention further comprises the following steps:
and calling the sample data stored in a distributed manner for machine learning, and constructing a prediction model.
The sample data refers to user data of the sample user and information characteristics of the sample information, and the click condition of the sample user on the sample information is known. The user data of the sample user and the information characteristics of the sample information can be stored in a plurality of recall servers in a scattered manner, so that the phenomenon that the sample data size is too large to influence the single machine operation is avoided. The server cluster can call sample data stored in a distributed mode in a plurality of recall servers to conduct machine learning, and a prediction model is built. For example, a large amount of sample data may be input into the logistic regression model, and parameters of the logistic regression model may be trained to obtain the prediction model.
On the basis of the above exemplary embodiment, as shown in fig. 7, the step 370 specifically includes:
in step 371, according to the historical browsing records of the user, all browsed information in the returned results is removed;
the historical browsing records refer to records left by the user in information browsing before summarizing the returned results. The history browsing records record records information browsed by the user. Specifically, each vertical service may remove browsed information from all returned results of a plurality of downstream information recall services according to a historical browsing record carried in user data.
In step 372, the remaining returned results are used to generate information recommendation results according to the configured adjustment strategy.
The configured adjustment policy may be to expand other interests than the interests recorded in the user data and obtain information related to the other interests. The configured adjustment strategy can also realize the diversity recommendation of information, and avoids recommending the information of the same type, the same field and the same person. Specifically, for each vertical service, information browsed by a user is removed according to information returned by a plurality of downstream information recall services, and according to a configured adjustment strategy, the remaining information is subjected to interest expansion, diversity screening and relevance ranking among user data to generate a final information recommendation result. And then, fusing the information recommendation results of all vertical services to obtain all information recommended to the user.
Fig. 8 is a schematic diagram of an architecture of a server cluster, which is exemplary of the present invention. As shown in fig. 8, the server cluster includes a proxy layer 801, a result fusion layer 802, a vertical service layer 803, and a recall layer 804. The proxy layer 801 receives a request carrying a user identifier sent by a mobile terminal. The result fusion layer 802 receives a request carrying a user identifier issued by the proxy layer 801 by running a Broker (scheduling) service, and acquires user data stored in correspondence with the user identifier. The user identification and the user data are stored in a key value pair mode. The result fusion layer 802 sends a request to the vertical service layer 803 to carry the user data. The result fusion layer 803 requests in parallel a plurality of information recall services of the downstream recall layer 804 to perform information recall by running a plurality of vertical services (teletext fusion service, video service, other service). Wherein each vertical service corresponds to a plurality of information recall services at the downstream. All the front-ranking data are segmented into a plurality of front-ranking data sets, and the recall layer 804 recalls the information of the front-ranking data sets configured by each information recall service by running all the information recall services. In the information recalling process, information which is not matched with the user data can be filtered, the click rate of the user on each piece of information is predicted by using the prediction model, and the final return result is determined by sequencing according to the click rate.
Each vertical service of the vertical service layer 803 may receive the returned results of the plurality of information recall services in the downstream recall layer 804, and perform history rearrangement (i.e., remove browsed information), diversity adjustment, interest expansion, and reordering on all the returned results. The result fusion layer 802 may perform fusion and aggregation on the reordered information of each vertical service of the vertical service layer to obtain an information recommendation result, and send the information recommendation result to the agent layer 801. Thereafter, the agent layer 801 sends the information recommendation result to the mobile terminal.
The server cluster divides all the front-row data into a plurality of front-row data sets, executes information recall on the front-row data sets configured by each information recall service through concurrence of a plurality of vertical services and concurrence of a plurality of information recall services at the downstream of each vertical service, collects the return results of all the information recall services according to a specified strategy, and collects the return results of all the vertical services to obtain an information recommendation result. Therefore, the processing process of information recommendation is evolved into parallel processing of distributed services, the data volume loaded by a single service is reduced, and the overall processing time is greatly reduced.
FIG. 9 is a block diagram of an architecture for performing information recommendation on a single machine. As shown in fig. 9, a single server includes a proxy layer and a recommendation system layer, where the proxy layer receives a request carrying a user identifier sent by a mobile terminal. And the recommendation system layer operates a recommendation service, acquires user data stored corresponding to the user identification, further executes information recall on all the currently-arranged data according to the user data, sorts the recall information according to the correlation between the user data and the recall information, and generates an information recommendation result according to specified strategies (such as weight arrangement, diversity, interest expansion and the like). And finally, returning the information recommendation result to the mobile terminal by the agent layer.
As can be seen from a comparison between fig. 8 and fig. 9, the information recall service of the present invention can utilize the information recall logic in a single machine of fig. 8, and can utilize a policy of the single machine to perform deduplication, interest expansion and information diversity on the returned results of all the information recall services. Therefore, the server cluster has low cost for transforming the single machine, but the data processing speed is greatly improved compared with the single machine.
Fig. 10 is a schematic diagram of an information recommendation system according to an exemplary embodiment of the present invention. As shown in fig. 10, the information recommendation system includes: a proxy server 1010, a scheduling server 1030, a fusion server 1050 and a plurality of recall servers 1070 connected in sequence;
the proxy server 1010 is used for receiving a request for information recommendation to a user;
as shown in fig. 10, the request may be sent by the mobile terminal 130 and carries the user identification of the sender. The proxy server 1010 receives a request for information recommendation to a user, which is sent by the mobile terminal 130.
The scheduling server 1030 is configured to obtain corresponding user data according to the user identifier indicated by the request;
the scheduling server 1030 stores the user data corresponding to each user identifier in a key value pair form, so that the scheduling server 1030 obtains the user data corresponding to the user identifier from its own storage according to the user identifier carried by the request.
The fusion server 1050 is used for requesting a plurality of recall servers 1070 to recall information in parallel through user data;
the fusion server 1050 includes one or more servers, and the fusion server 1050 sends an information recall request to the plurality of recall servers 1070 in parallel, where the information recall request carries user data and triggers the plurality of recall servers 1070 to run an information recall service deployed by itself for information recall.
The recall server 1070 is used for running an information recall service, the information recall service recommends and executes information recall for the user information from the configured positive data set according to the user data to obtain a return result of the information recall service, and the positive data set is formed by segmenting all positive data;
each recall server 1070 runs the deployed information recall service, and acquires information matched with the user data from the configured forward-arranged data set according to the user data carried by the information recall request to obtain a return result of the information recall service. Wherein all the front rank data are divided into a plurality of front rank data sets and are distributed and deployed in each recall server.
The scheduling server 1030 is further configured to aggregate the returned results of all the information recall services to form an information recommendation result.
The fusion server 1050 may first adjust the returned results of the plurality of recall servers 1070 downstream according to a specified policy (e.g., deduplication, diversity, interest expansion, etc.), and then the scheduling server 1030 may collect the returned results adjusted by the fusion server 1050 to obtain an information recommendation result. The scheduling server 1030 may also send the information recommendation result to the proxy server 1010, and the proxy server 1010 feeds back the information recommendation result to the mobile terminal 130.
The implementation process of the function and the role of each server in the information recommendation system may refer to the description in the embodiment of the information recommendation method, and is not described herein again.
The following is an embodiment of the apparatus of the present disclosure, which may be used to execute an embodiment of the information recommendation method executed by the server cluster of the present disclosure. For details not disclosed in the embodiments of the apparatus of the present disclosure, please refer to the embodiments of the information recommendation method of the present disclosure.
Fig. 11 is a block diagram illustrating an information recommendation apparatus according to an exemplary embodiment, which may be used in a server cluster of the implementation environment shown in fig. 1 to perform all or part of the steps of the information recommendation method shown in any one of fig. 3, 5 to 7. As shown in fig. 11, the apparatus includes, but is not limited to: a data acquisition module 1110, a parallel request module 1130, an information recall module 1150, and a result summarization module 1170. The information recall module 1150 is equivalent to the information recall service in the above method embodiment.
The data obtaining module 1110 is configured to obtain corresponding user data according to information recommendation triggered by a user;
a parallel request module 1130, configured to trigger information recall through a plurality of information recall modules deployed according to the acquired user data parallel request;
an information recall module 1150, configured to recommend and execute information recall for information of a user from a configured positive data set according to the user data, and obtain a return result of the information recall module, where the positive data set is formed by splitting all positive data;
and the result summarizing module 1170 is used for summarizing the returned results of all the information recalling modules to form an information recommendation result.
The implementation processes of the functions and actions of each module in the device are specifically described in the implementation processes of the corresponding steps in the information recommendation method, and are not described herein again.
The data acquisition module 1110 may be, for example, one of the physical structure input output interfaces 258 in fig. 2.
The parallel request module 1130, the information recall module 1150 and the result summarizing module 1170 may also be functional modules for performing corresponding steps in the information recommendation method. It is understood that these modules may be implemented in hardware, software, or a combination of both. When implemented in hardware, these modules may be implemented as one or more hardware modules, such as one or more application specific integrated circuits. When implemented in software, the modules may be implemented as one or more computer programs executing on one or more processors, such as programs stored in memory 232 for execution by central processor 218 of FIG. 2.
In an exemplary embodiment, the parallel request module 1130 includes:
and the data sending unit is used for sending the acquired user data to a plurality of deployed information recall modules in parallel, and the sending of the user data to the information recall modules triggers the information recall modules to execute information recall for the users.
In an exemplary embodiment, the information recommendation deploys a plurality of vertical services adapted to a plurality of recommendation information types, and a plurality of information recall modules are deployed downstream of the vertical services for information recall of corresponding recommendation information types; the data transmission unit includes:
and the service scheduling subunit is used for scheduling a plurality of vertical services and sending the user data to a plurality of information recall modules deployed at the downstream of each vertical service in parallel.
In an exemplary embodiment, as shown in fig. 12, the information recall module 1150 includes:
an information screening unit 1151, configured to screen, according to the user data, target information whose information characteristics match the user data from the configured forward data set;
a result obtaining unit 1152, configured to obtain a return result of the information recall module according to the target information.
In an exemplary embodiment, as shown in fig. 13, the result obtaining unit 1152 includes:
a click rate estimation subunit 11521, configured to calculate, by using the constructed prediction model, a click rate of the user to which the user data belongs to the target information;
and a click rate sorting subunit 11522, configured to sort the target information according to the size of the click rate, and determine a return result of the information recall module according to the sorting result.
Further, the information recommendation apparatus provided by the present invention further includes:
and the model construction module is used for calling the sample data of the distributed storage to perform machine learning and constructing the prediction model.
In an exemplary embodiment, as shown in fig. 14, the result summarization module 1170 comprises:
a history rearrangement unit 1171, configured to remove browsed information in all returned results according to the history browsing record of the user;
and a policy recommending unit 1172, configured to generate the information recommending result according to the configured adjustment policy from the remaining returned results.
Optionally, the present disclosure further provides an electronic device, where the electronic device may be used in a server cluster of the implementation environment shown in fig. 1 to execute all or part of the steps of the information recommendation method shown in any one of fig. 3 and 5 to 7. The electronic device includes:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to execute the information recommendation method of the above exemplary embodiment.
The specific manner in which the processor of the electronic device performs the operations in this embodiment has been described in detail in the embodiment related to the information recommendation method, and will not be elaborated herein.
In an exemplary embodiment, a storage medium is also provided that is a computer-readable storage medium, such as may be transitory and non-transitory computer-readable storage media, including instructions. The storage medium stores a computer program that can be executed by the central processors 222 of the plurality of servers 200 to perform the information recommendation method described above.
It will be understood that the invention is not limited to the precise arrangements 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 invention is limited only by the appended claims.
Claims (15)
1. An information recommendation method, comprising:
acquiring corresponding user data according to information recommendation triggered by a user;
triggering to recall information through a plurality of information recall services which are acquired and requested to be deployed by the user data in parallel; the information recall service is a program which is deployed on a server and used for executing information recall, and the information recall is a program for acquiring information which is possibly interested by a user;
the plurality of information recall services recommend and execute information recall for the information of the user from the respective configured main-line data sets according to the user data to obtain return results of the plurality of information recall services, the configured plurality of main-line data sets are formed by cutting all main-line data, and the main-line data is detailed data of the information;
and summarizing the returned results of all the information recall services to form an information recommendation result.
2. The method of claim 1, wherein the triggering of the recall of the information by the obtained plurality of information recall services deployed by the concurrent request of the user data comprises:
and sending the acquired user data to a plurality of deployed information recall services in parallel, wherein the sending of the user data to the information recall services triggers the information recall services to execute information recall for the users.
3. The method of claim 2, wherein the information recommendation deploys a plurality of vertical services adapted to a plurality of recommended information types, and a plurality of information recall services are deployed downstream of the vertical services for information recall of respective recommended information types;
the sending the acquired user data to a plurality of deployed information recall services in parallel comprises:
and sending the user data to a plurality of information recall services deployed downstream of each vertical service in parallel by scheduling a plurality of the vertical services.
4. The method of claim 1, wherein the plurality of information recall services respectively perform information recall for information recommendations of users from respective configured positive data sets according to the user data to obtain return results of the plurality of information recall services, and the method comprises:
the plurality of information recall services respectively screen out target information with information characteristics matched with the user data from the respective configured forward data sets according to the user data;
and obtaining the return results of the plurality of information recall services according to the target information.
5. The method of claim 4, wherein the obtaining the returned result of the information recall service according to the target information comprises:
the information recall service calculates the click rate of the user to which the user data belongs to the target information by using the established prediction model;
and sorting the target information according to the click rate, and determining a return result of the information recall service according to a sorting result.
6. The method of claim 5, wherein before the information recall service calculates the click rate of the user to which the user data belongs to the target information by using the constructed prediction model, the method further comprises:
and calling the sample data of the distributed storage to perform machine learning, and constructing the prediction model.
7. The method according to any one of claims 1 to 6, wherein the aggregating the returned results of all the information recall services to form an information recommendation result comprises:
removing browsed information in all returned results according to the historical browsing records of the user;
and generating the information recommendation result according to the rest return results and the configured adjustment strategy.
8. An information recommendation system, comprising: the system comprises a proxy server, a scheduling server, a fusion server and a plurality of recall servers which are sequentially connected;
the proxy server is used for receiving a request for information recommendation to a user;
the scheduling server is used for acquiring corresponding user data according to the user identification indicated by the request;
the fusion server is used for requesting a plurality of recall servers to recall information in parallel through the user data; the information recall service is a program which is deployed on a server and used for executing information recall, and the information recall is a program for acquiring information which is possibly interested by a user;
each recall server is used for operating information recall service, the information recall service recommends and executes information recall for information of a user from a configured front-ranking data set according to the user data to obtain a return result of the information recall service, the front-ranking data set is formed by segmenting all front-ranking data, and the front-ranking data are detail data of the information;
and the scheduling server is also used for summarizing the return results of all the information recall services to form an information recommendation result.
9. An information recommendation apparatus, comprising:
the data acquisition module is used for acquiring corresponding user data according to information recommendation triggered by the user;
the parallel request module is used for triggering and recalling information through a plurality of information recalling modules which are acquired and requested to be deployed by the user data in parallel; the information recall service is a program which is deployed on a server and used for executing information recall, and the information recall is a program for acquiring information which is possibly interested in a user;
each information recall module is used for recommending and executing information recall for the information of the user from the configured main data set according to the user data to obtain a return result of the information recall module, wherein the main data set is formed by segmenting all main data, and the main data is detailed data of the information;
and the result summarizing module is used for summarizing the returned results of all the information recalling modules to form an information recommendation result.
10. The apparatus of claim 9, wherein the parallel request module comprises:
and the data sending unit is used for sending the acquired user data to a plurality of deployed information recall modules in parallel, and the sending of the user data to the information recall modules triggers the information recall modules to execute information recall for the users.
11. The apparatus of claim 10, wherein the information recommendation deploys a plurality of vertical services adapted to a plurality of recommended information types, and a plurality of information recall modules are deployed downstream of the vertical services for information recall of respective recommended information types; the data transmission unit includes:
and the service scheduling subunit is used for scheduling a plurality of vertical services and sending the user data to a plurality of information recall modules deployed at the downstream of each vertical service in parallel.
12. The apparatus of claim 9, wherein the information recall module comprises:
the information screening unit is used for screening target information with information characteristics matched with the user data from the configured forward data set according to the user data;
and the result obtaining unit is used for obtaining the return result of the information recall module according to the target information.
13. The apparatus of claim 12, wherein the result obtaining unit comprises:
the click rate estimation subunit is used for calculating the click rate of the user to which the user data belongs to the target information by using the established prediction model;
and the click rate sorting subunit is used for sorting the target information according to the click rate and determining the returned result of the information recall module according to the sorting result.
14. An electronic device, characterized in that the electronic device comprises:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to perform performing the information recommendation method of any one of claims 1-7.
15. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program executable by a processor to perform the information recommendation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811020544.9A CN109408703B (en) | 2018-09-03 | 2018-09-03 | Information recommendation method and system, device, electronic equipment and storage medium thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811020544.9A CN109408703B (en) | 2018-09-03 | 2018-09-03 | Information recommendation method and system, device, electronic equipment and storage medium thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109408703A CN109408703A (en) | 2019-03-01 |
| CN109408703B true CN109408703B (en) | 2023-04-14 |
Family
ID=65463912
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811020544.9A Active CN109408703B (en) | 2018-09-03 | 2018-09-03 | Information recommendation method and system, device, electronic equipment and storage medium thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109408703B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110221817B (en) * | 2019-06-17 | 2023-01-17 | 北京酷我科技有限公司 | Data recall module and recommendation system |
| CN110633760B (en) * | 2019-09-25 | 2023-01-17 | 北京酷我科技有限公司 | Recommendation system integration method and recommendation system |
| CN110909200A (en) * | 2019-10-09 | 2020-03-24 | 上海麦克风文化传媒有限公司 | Offline and real-time combined audio album recommendation system recall method and system |
| CN110990695A (en) * | 2019-11-22 | 2020-04-10 | 厦门美柚股份有限公司 | Recommendation system content recall method and device |
| CN111190815B (en) * | 2019-12-19 | 2023-03-10 | 广州方硅信息技术有限公司 | Recall test method, apparatus, electronic device and storage medium |
| CN113032587B (en) * | 2019-12-25 | 2023-07-28 | 北京达佳互联信息技术有限公司 | Multimedia information recommendation method, system, device, terminal and server |
| CN113032588B (en) * | 2019-12-25 | 2024-03-08 | 北京达佳互联信息技术有限公司 | Multimedia recommendation system, method, device and equipment |
| CN111177563B (en) * | 2019-12-31 | 2023-06-27 | 北京顺丰同城科技有限公司 | Information recommendation method and device, electronic equipment and storage medium |
| CN111882399A (en) * | 2020-07-31 | 2020-11-03 | 平安国际融资租赁有限公司 | Service information recommendation method, device, computer system and readable storage medium |
| CN112182414A (en) * | 2020-08-13 | 2021-01-05 | 亿存(北京)信息科技有限公司 | Article recommendation method and device and electronic equipment |
| CN111966908A (en) * | 2020-08-25 | 2020-11-20 | 贝壳技术有限公司 | Recommendation system and method, electronic device, and computer-readable storage medium |
| CN113781146A (en) * | 2020-11-17 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Recommendation method, device, equipment and storage medium of product information |
| CN112765241A (en) * | 2021-02-04 | 2021-05-07 | 腾讯科技（深圳）有限公司 | Recall data determining method, apparatus and storage medium |
| CN113256393B (en) * | 2021-06-20 | 2021-10-29 | 浙江口碑网络技术有限公司 | Service information processing system, method and device and electronic equipment |
| CN113746913A (en) * | 2021-08-31 | 2021-12-03 | 上海明略人工智能(集团)有限公司 | Information recommendation method, system, device and storage medium |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103823805B (en) * | 2012-11-16 | 2018-10-19 | 腾讯科技（深圳）有限公司 | Community-based correlation note commending system and recommendation method |
| CN103778260A (en) * | 2014-03-03 | 2014-05-07 | 哈尔滨工业大学 | Individualized microblog information recommending system and method |
| CN106777094A (en) * | 2016-12-14 | 2017-05-31 | 大连大学 | The medical science big data searching system of the Skyline that view-based access control model vocabulary is matched with multiple features |
-
2018
- 2018-09-03 CN CN201811020544.9A patent/CN109408703B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109408703A (en) | 2019-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109408703B (en) | Information recommendation method and system, device, electronic equipment and storage medium thereof | |
| CN107423355B (en) | Information recommendation method and device, electronic equipment | |
| CN107222566B (en) | Information pushing method and device and server | |
| CN108763502B (en) | Information recommendation method and system | |
| CN105320766B (en) | Information-pushing method and device | |
| CN110543598B (en) | Information recommendation method and device and terminal | |
| CN110321422A (en) | Method, method for pushing, device and the equipment of on-line training model | |
| CN101266620B (en) | Method and equipment for providing target information for user | |
| US10262265B2 (en) | Systems and methods for generating and communicating application recommendations at uninstall time | |
| US20140280554A1 (en) | Method and system for dynamic discovery and adaptive crawling of content from the internet | |
| CN102346899A (en) | Method and device for predicting advertisement click rate based on user behaviors | |
| CN103329151A (en) | Recommendations based on topic clusters | |
| KR20130062442A (en) | Method and system for recommendation using style of collaborative filtering | |
| CN110704739A (en) | Resource recommendation method and device and computer storage medium | |
| CN110968767B (en) | Ranking engine training method and device, and business card ranking method and device | |
| CN108874812B (en) | Data processing method, server and computer storage medium | |
| CN109165975B (en) | Label recommending method, device, computer equipment and storage medium | |
| CN106156113A (en) | A kind of method of video recommendations, system and electronic equipment | |
| CN112560938B (en) | Model training method and device and computer equipment | |
| CN109271580B (en) | Search method, device, client and search engine | |
| CN112115364A (en) | Recommendation method and system for cold start based on login operation and computer equipment | |
| CN111275473B (en) | Content item delivery method, device, server and storage medium | |
| CN112182414A (en) | Article recommendation method and device and electronic equipment | |
| GB2448136A (en) | A content item distribution system | |
| CN112015924A (en) | Streaming media caching method and device and computer storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20221111 Address after: 1402, Floor 14, Block A, Haina Baichuan Headquarters Building, No. 6, Baoxing Road, Haibin Community, Xin'an Street, Bao'an District, Shenzhen, Guangdong 518000 Applicant after: Shenzhen Yayue Technology Co.,Ltd. Address before: 518000 Tencent Building, No. 1 High-tech Zone, Nanshan District, Shenzhen City, Guangdong Province, 35 Floors Applicant before: TENCENT TECHNOLOGY (SHENZHEN) Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |