Content recommending method and device based on user characteristics

[0032] The present invention will be described in further detail below in conjunction with the accompanying drawings.

[0033] figure 1 Shows a system topology diagram for content recommendation based on user characteristics according to one aspect of the present invention, which includes a network device 1 and multiple user devices 2. The network device 1 may be connected to multiple user devices 2 via a network, and recommend content for the user according to the user access behavior initiated by the user from the user device 2. Here, the network includes but not limited to the Internet, WAN, metropolitan area network, local area network, VPN network, wireless ad hoc network (AdHoc network), etc.; user access behavior includes but not limited to the user's information browsing and information search through the user device 2 , Operation behavior during activities such as interacting with friends and shopping; the content recommended by the network device 1 for the user includes but not limited to information resources such as network information, network services, online shopping, online games, and online communities.

[0034] In addition, the network device 1 may be an independent computing device, including but not limited to ordinary computers, servers, small hosts, mainframes, etc., or may be a computing device shared with other applications, including but not limited to enterprise servers, enterprise website servers, Portal server, search server, etc. The user equipment 2 can be any electronic product that can interact with the user through a keyboard, mouse, remote control, touchpad, or handwriting device, such as computers, mobile phones, PDAs, game consoles, handheld computers, PPCs, Tablet PC, or IPTV, etc.

[0035] Those skilled in the art should understand that the above-mentioned network equipment 1, user equipment 2 and the network connected between them are only examples, and other existing or possible future network equipment, user equipment or networks, if applicable to the present invention, should also include It is within the protection scope of the present invention and is included here by reference.

[0036] For the sake of brevity, a user equipment 2 is used as an example for description. Those skilled in the art should understand that the network device 1 can interact with multiple user devices 2 at the same time, and recommend content for users according to user access behaviors initiated from different user devices 2.

[0037] figure 2 A schematic diagram showing a network device for content recommendation based on user characteristics according to one aspect of the present invention. Among them, the network device 1 includes an acquisition device 11, a classification device 12, a mining device 13, and a recommendation device 14.

[0038] Specifically, the obtaining means 11 obtains the user access behavior of the user from the user equipment 2. That is, the user initiates user access behavior through user equipment 2, for example, accesses the Internet through browser software or client software in user equipment 2, and interacts with user equipment 2, including but not limited to keyboard, mouse, and remote control. , Touchpad, or handwriting device to perform network access operations such as information browsing, searching, and instant messaging with friends; taking the keyboard as an example, when the user taps the keyboard button in the search bar of the search page in the browser of the user device 2 When inputting and clicking Enter to submit an information search request, the obtaining device 11 obtains relevant field information from the information search request sent from the browser to the server in real time according to the communication protocol used by the browser, such as http or https. , And record as the user's user access behavior. Preferably, the obtaining device 11 records and merges user access behaviors initiated by the user through different user equipment 2 according to the user name (login ID) used by the user to log in to the Internet or network application. More preferably, the obtaining means 11 uses the conversation process between the user and the network application that he accesses to record and merge the user access behavior initiated by the user through the user equipment 2 during the validity period of the process. More preferably, the acquiring device 11 may use a random code generating device to require the user to fill in a randomly generated random code when initiating user access behavior, so as to prevent robots from publishing information or access behavior initiated by viruses such as Trojan horses, thereby improving network equipment 1 Obtain the accuracy of user access behavior. Preferably, the network device 1 may include a user access behavior library 101 to store the user access behaviors acquired by the acquiring device 11; the user access behavior library 101 can not only provide user access behavior data required for a longer-time granular user access behavior analysis, It can also alleviate the "producer-consumer" matching contradiction between the acquisition device 11 and the classification device 12 to a certain extent.

[0039] The classification device 12 classifies users according to the user access behavior provided by the acquisition device 11 to obtain one or more user groups. Specifically, the classification device 12 analyzes and processes the user access behavior, for example, extracts information such as the terminal type of the user equipment, the location of the user, the application service accessed by the user, and the time period the user accesses based on the user access behavior, and According to this information, these users are classified to obtain one or more user groups, such as the user group using a certain terminal type of user equipment 2, the user group accessing a certain application service, and the user group in a certain location. Wait. Those skilled in the art should understand that the above-mentioned methods for classifying users are only examples. Other existing or future user classification methods that are applicable to the present invention should also be included in the scope of protection of the present invention. Included here by reference. Preferably, the classification device 12 can use a variety of methods to classify users according to the user's access behavior to obtain one or more user groups. The multiple methods include, but are not limited to: for each user access behavior received by the classification device 12 in real time and acquired by the acquisition device 11 in real time, immediately update the user and the user group corresponding to the user access behavior; or for classification The user access behavior received by the device 12 in real time and acquired by the acquiring device 11 over a period of time, and the user group corresponding to each of the user and the user access behavior is updated one by one until all the users in the user access behavior The behaviors have been processed; or for the user access behaviors received by the classification device 12 and acquired by the acquisition device 11 within a period of time, perform statistical analysis and processing, and update the user and the user's access behavior corresponding to the results of the statistical analysis. user group. Those skilled in the art should understand that the above-mentioned methods of determining user group membership from user access behaviors are only examples. Appropriate changes based on these examples can also be applied to the present invention. The difference may only lie in the difference between the user and the user group to which it belongs. The changes are different, so they should also be included in the protection scope of the present invention, and are included here by reference. Preferably, the network device 1 may include a user classification database 102 to store users and user groups to which they belong; the user classification database 102 updates users and user groups to which they belong based on the information provided by the classification device 12, and provides the network device 1 with the necessary information. User group information.

[0040] The mining device 13 determines the information resource matching the user group according to the user group provided by the classification device 12. Specifically, the mining device 13 performs data mining on the user group obtained by dividing the users by the classification device 12, such as performing statistical research and analysis on the user access behavior of users in the user group to determine information resources suitable for the user group. For example, the mining device 13 conducts statistical research on the user access behaviors of users in a user group, and finds that users in the user group who exceed a certain percentage of a threshold are interested in a certain information resource, and will infer that other users in the user group should also respond to it. An information resource is interested, so that this information resource is determined as an information resource that matches the user group. For example, if the mining device 13 finds that more than half of the users in a regional user group (such as "Shanghai Xujiahui User Group") are using "Dianping.com (www.dianping.com)", it will infer the "Shanghai Xujiahui User Group" Other users are also interested in this "Dianping.com". Therefore, this "Dianping.com" will serve as one of the information resources that matches the "Shanghai Xujiahui User Group". Preferably, the network device 1 may include an access resource library 103 to store information resources that each user in the user group has accessed; the access resource library 103 updates each user in the user group according to user access behavior and user group information provided by the classification device 12 Information resources that have been accessed, and provide the mining device 13 with required access resource information.

[0041] The recommending device 14 recommends the information resource to the user equipment of at least one user in the user group according to the information resource provided by the mining device 13. Specifically, the recommending device 14 recommends the relevant information of the information resource based on the information resource matching the user group provided by the mining device 13, such as the information resource "Dianping.com" that matches the "Shanghai Xujiahui User Group" in the previous example One or more users in the user group, that is, the user equipment 2 recommended to these target users, for the target users to browse. The way that the recommending device 14 recommends the relevant information of the matching information resources to one or more users in the user group, including but not limited to: pushing the required recommended content to the user equipment used by these target users to initiate user access behaviors 2; Or use page technology, such as JSP, ASP, PHP technology, write the required content to one or more web pages, and push the URL of the relevant web page to the user equipment used by these target users to initiate user access behavior 2 ; Or use page technology, such as JSP, ASP, PHP technology, write the content to be promoted into one or more web pages, and push part of the content to be promoted and the URL of related web pages to these target users to initiate user access behavior User equipment used 2. Among them, for the latter two referral methods, if the user needs to know the content to be recommended or other content that is not displayed, he can directly access the webpage corresponding to the URL provided by the referral device 14 through his user equipment 2. Those skilled in the art should understand that the above-mentioned page technologies are only examples, and other existing or future page technologies that are applicable to the present invention should also be included in the scope of protection of the present invention and are cited here. Included here. Similarly, those skilled in the art should understand that the above-mentioned several promotion methods are only examples, and other existing or future promotion methods that are applicable to the present invention should also be included in the scope of protection of the present invention, and are used here. The citation method is included here. Preferably, the network device 1 may include an information resource library 104 to store relevant information about information resources to be recommended to related users in each user group; the information resource library 104 is updated according to the information resource information of each user group provided by the mining device 13 Information about the information resources to be recommended to related users in each user group is needed, and the recommending device 14 is provided with required information resource information.

[0042] Preferably, the acquiring device 11, the sorting device 12, the excavating device 13, and the recommending device 14 work continuously. Specifically, the acquiring device 11 acquires the user access behavior of the user from the user equipment 2 in real time or periodically; the classification device 12 classifies the user according to the acquired user access behavior to obtain one or more user groups; the mining device 13 The user group determines the information resource that matches the user group; and the recommending device 14 recommends the information resource to the user equipment of at least one user in the user group. Here, those skilled in the art should understand that “continuous” means that each device of the network device 1 performs user access behavior acquisition, user classification, determination of user group information resources, and user group information according to the set or real-time adjustment work mode requirements. Referral until the user stops performing network operations through the user equipment 2 for a long time, that is, the user access behavior.

[0043] In a preferred embodiment, the classification device 12 is further configured to perform clustering processing on at least one of the one or more user groups to obtain one or more user groups and their clustering characteristics. Specifically, for a user group, the classification device 12 first obtains the description information of each user in the user group, such as a user characterization vector, and then uses a clustering algorithm, such as k-means, k-center point algorithm, to determine the user group Decompose into several new user groups, and obtain the clustering characteristics of each new user group, so that in the sense of clustering characteristics, each user within these new user groups has a higher degree of similarity, and each user between the groups Have a higher degree of difference. For example, the classification device 12 extracts keywords from the user access behavior of each user to obtain multiple keywords and their frequencies, thereby constituting the user characterization vector of the user; then, the vector distance algorithm is used to obtain the information between different users. Vector distance; Finally, users with smaller vector distances are combined into a new user group, and the user characterization vector of a user in the new user group or the average of the user characterization vectors of all users is taken as the clustering feature of the user group . For example, for the "Shanghai Xujiahui user group", the classification device 12 extracts keywords from the user's access behavior in the user group to obtain a user characterization vector, and then applies a clustering algorithm to obtain multiple new user groups, such as clustering features. The "Shanghai Xujiahui Wine User Group" with "Red Wine" and the "Shanghai Xujiahui Tourist User Group" with the clustering feature of "Tourism". Those skilled in the art should understand that the above-mentioned clustering algorithms are only examples, and other existing or future clustering algorithms that are applicable to the present invention should also be included in the scope of protection of the present invention, and are used here The citation method is included here.

[0044] In another preferred embodiment, the mining device 13 is further configured to determine information resources matching the user group based on the user group and its clustering characteristics. Specifically, for a user group with a certain clustering feature, the mining device 13 automatically obtains related information resources matching the user group according to its clustering feature, such as using the clustering feature as a keyword to obtain related information through a search engine. Information resources, or matching search related updated information resources. Following the example above, for the "Shanghai Xujiahui Wine User Group" whose clustering feature is "wine", the mining device 13 obtains information resources related to "wine" through a search engine, such as introduction to wine appreciation, wine and health knowledge, or matching search To update information resources related to "wine", such as wine tasting activities, as an information resource that matches the "Shanghai Xujiahui Wine User Group". Those skilled in the art should understand that the above-mentioned algorithms for determining information resources based on clustering characteristics are only examples. Other existing or future algorithms for determining information resources based on clustering characteristics should also be applicable to the present invention. It is included in the protection scope of the present invention and is included here by reference.

[0045] Preferably, the recommending device 14 is further configured to recommend information resources matching the user equipment to the user equipment of at least one user in the user group. Specifically, for at least one user in the user group, the recommending device 14 determines the terminal attributes of the user equipment 2 used by the user according to the user access behaviors of these target users, and converts the information content of the information resource into its terminal The format that matches the attributes, or the format that matches the attributes of the terminal is searched for, and the information content is promoted to the user device 2 for the target user to browse. For example, when the recommending device 14 needs to recommend a picture of "Shrek" to the user, the recommending device 14 determines the terminal attributes of the user equipment 2 used by the user according to the user access behavior of these target users, such as resolution_width, resolution_height, search The "Shrek" pictures that best match the resolution, or the "Shrek" pictures whose resolution exceeds the resolution attribute of the user device 2 are converted into a format suitable for the user device 2, and these pictures are promoted to the user Device 2, for the user to browse. Those skilled in the art should understand that the above-mentioned adaptation methods for providing information resources matching the user equipment 2 are only examples, and other existing or future adaptation methods that may be applicable to the present invention should also include It is within the protection scope of the present invention and is included here by reference.

[0046] More preferably, the classification device 12 is further configured to determine and/or update the user characteristics of the user according to the user's access behavior; according to the user characteristics, to classify the user to obtain one or more users group. Specifically, the classification device 12 extracts corresponding fields from the access request corresponding to the user's access behavior according to the communication protocol between the user equipment 2 and the network device 1, such as http or https, to obtain and/or update the user characteristics of the user; These user characteristics classify these users to obtain one or more user groups. Here, the user characteristics include, but are not limited to, characteristic information such as the terminal type of the user equipment 2, the location of the user, the application service accessed by the user, and the time period during which the user accesses. Those skilled in the art should understand that the above-mentioned user features are only examples, and other existing or future user features that are applicable to the present invention should also be included in the scope of protection of the present invention and are cited here. Included here. For example, the classification device 12 determines that the user characteristics of these users include location characteristics such as "Shanghai Xujiahui", "Beijing Haidian District", and "Guangzhou Tianhe District" according to the user's access behavior, and classifies these users according to these user characteristics to obtain Multiple user groups, such as "Shanghai Xujiahui User Group", "Beijing Haidian District User Group", "Guangzhou Tianhe District User Group", etc.

[0047] image 3 It shows a schematic diagram of a network device for content recommendation based on user characteristics according to a preferred embodiment of the present invention. Among them, the network device 1 further includes a frequency adjustment device 15. Specifically, the frequency adjustment device 15 adjusts the acquisition frequency for acquiring the user's access behavior according to the changes of the user and the user group to which it belongs; and feeds back the adjusted acquisition frequency to the acquisition device 11 for it to acquire the information. Describe user access behavior. Preferably, the frequency adjustment device accumulates changes in the user and the user group to which they belong over a period of time, and when the accumulated value is lower than a set frequency reduction threshold, the acquisition frequency is reduced; when the accumulated value When the set frequency up threshold is exceeded, the acquisition frequency is increased. Here, the change between a user and the user group to which it belongs means that the user has left or joined a user group. For example, if the user changes the user group to which he belongs, and uses one user group to join another user group, then the user and his The cumulative number of changes in the user group is twice.

[0048] Figure 4 A schematic diagram showing a network device and a user device for content recommendation based on user characteristics according to another aspect of the present invention. The user equipment 2 includes an access device 21, and the network device 1 includes an acquisition device 11, a classification device 12, a mining device 13, and a recommendation device 14. Specifically, the access device 21 initiates the user access behavior carrying the user mark from the user equipment 2, and the acquisition device 11 obtains the user access behavior and the user mark from the user equipment 2 (that is, initiated by the access device 21). That is, the access device 21 loads the user tag corresponding to the user in the agreed field in the access request corresponding to the user's access behavior, such as the protocol header of http or https or a certain data segment of the data unit; The access behavior of the user side is described, and the user mark of the user is extracted from the corresponding agreement field of the access request corresponding to the user access behavior. Further, in the process of classifying users according to user access behaviors, the classification device 12 preferably combines user access behaviors with the same user tag, and then performs classification processing, so as to obtain better classification and content recommendation effects. Since the classification process in the classification device 12, the excavating device 13 and the recommending device 14 are basically the same as those in the previous embodiment, they will not be repeated here, and they are included here by reference.

[0049] Figure 5 A schematic diagram showing a network device and a user device for content recommendation based on user characteristics according to another preferred embodiment of the present invention. In this embodiment, the user equipment 2 further includes a second marking device 22, and the network equipment 1 further includes a first marking device 16. Specifically, the second marking device 22 sends the device identification code of the user equipment 2 to the network device 1; the first marking device 16 receives the device identification code of the user equipment sent from the user equipment 2, and detects a new user based on the device identification code. The device generates a corresponding user mark for it, and sends the user mark to the user equipment 2; the second marking device 22 receives the user mark of the user equipment 2 sent by the network device 1. For example, for a user who uses a mobile terminal as the user equipment 2, the second marking device 22 sends the IMEI (International Mobile Equipment Identification) of the user equipment 2 as the equipment identification code to the network equipment 1; After receiving the IMEI of the user equipment 2, the device 16 detects whether the IMEI already exists, and if it does not exist (indicating that the user equipment 2 is a new user equipment), it generates a user mark for the user equipment 2 and marks the user Sent to the user equipment 2; the second marking device 22 receives the user marking sent by the network equipment 2. Here, the device identification code can be the hardware identification code of the user equipment 2, such as the serial number of the complete machine or the serial number of a certain component, or the system software identification code of the user equipment 2, such as the serial number of the system software, or It is the combination of the software and hardware identification codes of the user equipment 2. Those skilled in the art should understand that the above-mentioned equipment identification codes are only examples, and other existing or future equipment identification codes that are applicable to the present invention should also be included in the scope of protection of the present invention, and are used here. The citation method is included here.

[0050] Subsequently, in this embodiment, the access device 21 initiates the user access behavior carrying the user mark from the user equipment 2, and the acquisition device 11 acquires the user access behavior of the user from the user equipment 2 (that is, initiated by the access device 21) and the User tag. That is, the access device 21 loads the user tag corresponding to the user in the agreed field in the access request corresponding to the user's access behavior, such as the protocol header of http or https or a certain data segment of the data unit; The access behavior of the user side is described, and the user mark of the user is extracted from the corresponding agreement field of the access request corresponding to the user access behavior. Preferably, the network device 1 further includes a user identification library 106 to store user identification and other information, such as the correspondence between the equipment identification code and the user identification; the user identification library 106 is based on the correspondence between the equipment identification code of the first identification device 16 and the user identification. The relationship updates the relevant content of the user ID library for further processing.

[0051] Image 6 A flowchart of a method according to one aspect of the present invention is shown, which shows a process of content recommendation based on user characteristics.

[0052] Specifically, in step S1, the network device 1 obtains the user's user access behavior from the user device 2. That is, the user initiates user access behavior through user equipment 2, for example, accesses the Internet through browser software or client software in user equipment 2, and interacts with user equipment 2, including but not limited to keyboard, mouse, and remote control. , Touchpad, or handwriting device to perform network access operations such as information browsing, searching, and instant messaging with friends; taking the keyboard as an example, when the user taps the keyboard button in the search bar of the search page in the browser of the user device 2 When inputting and clicking Enter to submit the information search request, in step S1, the network device 1 sends the information search request from the browser to the server in real time according to the communication protocol used by the browser, such as http or https Obtain the relevant field information and record the user access behavior of the user. Preferably, in step S1, the network device 1 records and merges user access behaviors initiated by the user through different user devices 2 according to the user name (login ID) used by the user to log in to the Internet or the network application. More preferably, in step S1, the network device 1 uses the session process between the user and the network application accessed by the user to record and merge the user access behavior initiated by the user device 2 during the validity period of the process. More preferably, in step S1, the network device 1 may use the random code generating device to request the user to fill in a randomly generated random code when initiating user access behavior, so as to prevent robots from publishing information or access behavior initiated by viruses such as Trojan horses. Thus, the accuracy of the network device 1 in obtaining the user's access behavior is improved. Preferably, the network device 1 may include a user access behavior library 101 to store the user access behaviors obtained in step S1; the user access behavior library 101 can not only provide the required user access behaviors for longer-time granular user access behavior analysis The data can also alleviate the "producer-consumer" matching contradiction between step S1 and step S2 of the user equipment 1 to a certain extent.

[0053] In step S2, the network device 1 classifies users according to the user access behavior provided in step S1 to obtain one or more user groups. Specifically, in step S2, the network device 1 analyzes and processes the user access behavior, for example, extracts the terminal type of the user equipment, the location of the user, the application service accessed by the user, and the time of the user access based on the user access behavior. According to this information, these users are classified to obtain one or more user groups, such as user groups using a certain terminal type of user equipment 2, user groups accessing a certain application service, User groups in the location area, etc. Those skilled in the art should understand that the above-mentioned methods for classifying users are only examples. Other existing or future user classification methods that are applicable to the present invention should also be included in the scope of protection of the present invention. Included here by reference. Preferably, in step S2, the network device 1 may use multiple methods to classify users according to the user access behavior, so as to obtain one or more user groups. The multiple methods include, but are not limited to: for each user access behavior received by the network device 1 in real time in step S2 and obtained in real time in step S1 by the network device 1, immediately update the user and the user access behavior corresponding to the user. User group; or for the user access behavior received by the network device 1 in step S2 in real time and obtained in step S1 over a period of time, update the user group corresponding to each of the user access behaviors one by one , Until all the user access behaviors in the user access behavior have been processed; or for the user access behavior received by the network device 1 in step S2 and obtained in step S1 within a period of time, perform statistical analysis and processing , And update the user group corresponding to the user and the user's access behavior based on the statistical analysis result. Those skilled in the art should understand that the above-mentioned methods of determining user group membership from user access behaviors are only examples. Appropriate changes based on these examples can also be applied to the present invention. The difference may only lie in the difference between the user and the user group to which it belongs. The changes are different, so they should also be included in the protection scope of the present invention, and are included here by reference. Preferably, the network device 1 may include a user classification database 102 to store users and user groups to which they belong; the user classification database 102 updates the users and user groups to which they belong based on the information provided in step S2, and provides the network device 1 with Required user group information.

[0054] In step S3, the network device 1 determines an information resource matching the user group according to the user group provided in step S2. Specifically, in step S3, the network device 1 performs data mining on the user group obtained by dividing users in step S2, such as conducting statistical research and analysis on the user access behavior of users in the user group to determine suitable Information resources for this user group. For example, in step S3, the network device 1 conducts statistical research on the user access behaviors of users in a user group, and finds that users in the user group exceeding a certain percentage of the threshold are interested in a certain information resource, and will infer the information resources in the user group. Other users should also be interested in this information resource, so that this information resource is determined as an information resource that matches the user group. For example, in step S3, network device 1 finds that more than half of users in a regional user group (such as "Shanghai Xujiahui User Group") are using "Dianping Network (www.dianping.com)", then it will infer "Shanghai Other users of "Xujiahui User Group" are also interested in this "Dianping Network". Therefore, this "Dianping Network" will be used as one of the information resources matching the "Shanghai Xujiahui User Group". Preferably, the network device 1 may include an access resource library 103 to store the information resources that each user in the user group has accessed; the access resource library 103 updates the user group according to the user access behavior and the user group information provided in step S2 The information resources that each user has accessed, and the required access resource information is provided for step S3.

[0055] In step S4, the network device 1 recommends the information resource to the user equipment of at least one user in the user group according to the information resource provided in step S3. Specifically, in step S4, the network device 1 uses the information resource matching the user group provided by it in step S3, such as the information resource "Dianpingwang" that matches the "Shanghai Xujiahui User Group" in the previous example, to The relevant information of the information resource is recommended to one or more users in the user group, that is, to the user equipment 2 of these target users, for the target users to browse. In step S4, the network device 1 recommends the related information of the matching information resource to one or more users in the user group, including but not limited to: pushing the content to be recommended to these target users to initiate user access behavior The user equipment used 2; or use page technology, such as JSP, ASP, PHP technology, write the required content to one or more web pages, and push the URL of the relevant web page to these target users to initiate user access behavior User device 2 used; or use page technologies, such as JSP, ASP, and PHP technologies, to write the content to be promoted into one or more web pages, and push part of the content to be promoted and the URLs of related web pages to these targets User equipment 2 used by the user to initiate user access. Among them, for the latter two referral methods, if the user needs to know the content to be promoted or other content that is not displayed, he can directly access the URL corresponding to the URL provided by the network device 1 in step S4 through his user device 2 Page. Those skilled in the art should understand that the above-mentioned page technologies are only examples, and other existing or future page technologies that are applicable to the present invention should also be included in the scope of protection of the present invention and are cited here. Included here. Similarly, those skilled in the art should understand that the above-mentioned several promotion methods are only examples, and other existing or future promotion methods that are applicable to the present invention should also be included in the scope of protection of the present invention, and are used here. The citation method is included here. Preferably, the network device 1 may include an information resource library 104 to store related information about information resources to be recommended to related users in each user group; the information resource library 104 is based on the information of each user group provided by the network device 1 in step S3. The information resource information updates the relevant information of the information resource to be recommended to the relevant users in each user group, and provides the required information resource information for step S4.

[0056] Preferably, steps S1, S2, S3 and S4 work continuously. Specifically, in step S1, the network device 1 obtains the user's user access behavior from the user device 2 in real time or periodically; in step S2, the user is classified according to the obtained user access behavior to obtain one or more users Group; In step S3, according to the user group, determine the information resource matching the user group; and in step S4, recommend the information resource to the user equipment of at least one user in the user group. Here, those skilled in the art should understand that “continuous” means that each step of the network device 1 performs user access behavior acquisition, user classification, and determination of user group information resources in accordance with the set or real-time adjustment work mode requirements. Referral until the user stops performing network operations through the user equipment 2 for a long time, that is, the user access behavior.

[0057] In a preferred embodiment, in step S2, the network device 1 is further configured to perform clustering processing on at least one of the one or more user groups to obtain one or more user groups and their clustering characteristics. Specifically, for a user group, in step S2, the network device 1 first obtains the description information of each user in the user group, such as a user characterization vector, and then uses a clustering algorithm, such as k-mean, k-center point algorithm , Decompose the user group into several new user groups, and obtain the clustering characteristics of each new user group, so that in the sense of clustering characteristics, each user within these new user groups has a higher degree of similarity. Each user between has a higher degree of difference. For example, in step S2, the network device 1 extracts keywords from the user access behavior of each user to obtain multiple keywords and their frequencies to form a user representation vector of the user; then, the vector distance algorithm is used to obtain The vector distance between different users; finally, the users with the smaller vector distance are combined into a new user group, and the user characterization vector of a user in the new user group or the average of the user characterization vectors of all users is taken as the user Clustering characteristics of the group. For example, for the "Shanghai Xujiahui user group", in step S2, the network device 1 extracts keywords from the user's access behavior in the user group to obtain the user representation vector, and then uses the clustering algorithm to obtain multiple new user groups , Such as the "Shanghai Xujiahui Red Wine User Group" with the clustering feature of "wine" and the "Shanghai Xujiahui Tourist User Group" with the clustering feature of "tourism". Those skilled in the art should understand that the above-mentioned clustering algorithms are only examples, and other existing or future clustering algorithms that are applicable to the present invention should also be included in the scope of protection of the present invention, and are used here The citation method is included here.

[0058] In another preferred embodiment, in step S3, the network device 1 is further configured to determine information resources that match the user group according to the user group and its clustering characteristics. Specifically, for a user group with a certain clustering feature, in step S3, the network device 1 automatically obtains related information resources that match the user group according to the clustering feature, such as using the clustering feature as a keyword Get related information resources through search engines, or match and retrieve related updated information resources. Following the above example, for the "Shanghai Xujiahui Wine User Group" whose clustering feature is "wine", in step S3, network device 1 obtains information resources related to "wine" through a search engine, such as introduction to wine appreciation, wine and health Knowledge, or matching retrieval of updated information resources related to "wine", such as wine tasting activities, as an information resource that matches the "Shanghai Xujiahui Wine User Group". Those skilled in the art should understand that the above-mentioned algorithms for determining information resources based on clustering characteristics are only examples. Other existing or future algorithms for determining information resources based on clustering characteristics should also be applicable to the present invention. It is included in the protection scope of the present invention and is included here by reference.

[0059] Preferably, in step S4, the network device 1 is further configured to recommend information resources matching the user equipment to the user equipment of at least one user in the user group. Specifically, for at least one user in the user group, in step S4, the network device 1 determines the terminal attributes of the user device 2 used by the user according to the user access behaviors of these target users, and compares the information of the information resource The content is converted into a format that matches the attributes of the terminal, or a format that matches the attributes of the terminal is searched for, and the information content is promoted to the user device 2 for the target user to browse. For example, when in step S4, the network device 1 needs to recommend a picture of "Shrek" to the user, in step S4, the network device 1 determines the terminal of the user device 2 used by the user according to the user access behavior of these target users Properties, such as resolution_width, resolution_height, search for the "Shrek" picture that best matches the resolution, or convert the "Shrek" picture whose resolution exceeds the resolution property of the user device 2 into a suitable one for the user device 2. Format and recommend these pictures to the user’s device 2 for the user to browse. Those skilled in the art should understand that the above-mentioned adaptation methods for providing information resources matching the user equipment 2 are only examples, and other existing or future adaptation methods that may be applicable to the present invention should also include It is within the protection scope of the present invention and is included here by reference.

[0060] More preferably, in step S2, the network device 1 is further configured to determine and/or update the user characteristics of the user according to the user access behavior; according to the user characteristics, perform classification processing on the user to obtain One or more user groups. Specifically, in step S2, the network device 1 extracts corresponding fields from the access request corresponding to the user's access behavior according to the communication protocol between the user equipment 2 and the network device 1, such as http or https, to obtain and/or update the user's User characteristics; according to these user characteristics, these users are classified to obtain one or more user groups. Here, the user characteristics include, but are not limited to, characteristic information such as the terminal type of the user equipment 2, the location of the user, the application service accessed by the user, and the time period during which the user accesses. Those skilled in the art should understand that the above-mentioned user features are only examples, and other existing or future user features that are applicable to the present invention should also be included in the scope of protection of the present invention and are cited here. Included here. For example, in step S2, the network device 1 determines the user characteristics of these users including location characteristics such as "Shanghai Xujiahui", "Beijing Haidian District", and "Guangzhou Tianhe District" based on the user's access behavior, and based on these user characteristics, the user Perform classification processing to obtain multiple user groups, such as "Shanghai Xujiahui User Group", "Beijing Haidian District User Group", "Guangzhou Tianhe District User Group", etc.

[0061] Figure 7 A flowchart of a method according to a preferred embodiment of the present invention is shown, which shows a process of content recommendation based on user characteristics. Specifically, in step S5, the network device 1 adjusts the acquisition frequency for acquiring the user's access behavior according to changes in the user and the user group to which it belongs; and feeds back the adjusted acquisition frequency to step S1 for It obtains the user access behavior. Preferably, the frequency adjustment device accumulates changes in the user and the user group to which they belong over a period of time, and when the accumulated value is lower than a set frequency reduction threshold, the acquisition frequency is reduced; when the accumulated value When the set frequency up threshold is exceeded, the acquisition frequency is increased. Here, the change between a user and the user group to which it belongs means that the user has left or joined a user group. For example, if the user changes the user group to which he belongs, and uses one user group to join another user group, then the user and his The cumulative number of changes in the user group is twice.

[0062] Figure 8 A flowchart of a method according to another aspect of the present invention is shown, which shows a process in which a network device and a user device cooperate to implement content recommendation based on user characteristics. Specifically, in step s1, the user equipment 2 initiates a user access behavior carrying a user mark, and the network device 1 obtains the user access behavior and the user mark of the user from the user equipment 2. That is, in step s1, the user equipment 2 loads the user tag corresponding to the user in the agreed field in the access request corresponding to its user access behavior, such as the http or https protocol header or a certain data segment of the data unit; The network device 1 obtains the user-side access behavior, and extracts the user mark of the user in the corresponding appointment field of the access request corresponding to the user access behavior. Further, in step s2, in the process of classifying users according to user access behaviors, the network device 1 preferably combines user access behaviors with the same user tag, and then performs classification processing, so as to obtain better classification and content. Recommended effect. Since the classification process, step s3, and step s4 of the network device 1 in step s22 are basically the same as the steps S2, S3, and S4 in the foregoing embodiment, they will not be described in detail, and are included here by reference.

[0063] Picture 9 A flowchart of a method according to another preferred embodiment of the present invention is shown, which shows a process in which network equipment and user equipment cooperate to implement content recommendation based on user characteristics. Specifically, in step s6, user equipment 2 sends the device identification code of user equipment 2 to network device 1; network device 1 receives the device identification code of the user equipment sent from user equipment 2; in step s7, network device 1 The new user equipment is detected according to the device identification code and a corresponding user mark is generated for it; in step s8, the network device 1 sends the user mark to the user equipment 2, and the user equipment 2 receives the information about the user sent by the network device 1 User tag for device 2. For example, for a user who uses a mobile terminal as the user equipment 2, in step s6, the user equipment 2 sends the IMEI (International Mobile Equipment Identification) of the user equipment 2 as the equipment identification code to the network equipment 1; In step s7, after receiving the IMEI of the user equipment 2, the network equipment 1 detects whether the IMEI already exists, and if it does not exist (indicating that the user equipment 2 is a new user equipment), it generates a user tag for the user equipment 2 In step s8, the network device 1 sends the user mark to the user device 2, and the user device 2 receives the user mark sent by the network device 2. Here, the device identification code can be the hardware identification code of the user equipment 2, such as the serial number of the complete machine or the serial number of a certain component, or the system software identification code of the user equipment 2, such as the serial number of the system software, or It is the combination of the software and hardware identification codes of the user equipment 2. Those skilled in the art should understand that the above-mentioned device identification codes are only examples, and other existing or future device identification codes that are applicable to the present invention should also be included in the scope of protection of the present invention, and are used here The citation method is included here.

[0064] Subsequently, in this embodiment, in step s1, the user equipment 2 initiates the user access behavior carrying the user mark from the user equipment 2, and the network equipment 1 obtains the user equipment 2 (that is, initiated by the user equipment 2) The user's access behavior is marked with the user. That is, in step s1, the user equipment 2 loads the user tag corresponding to the user in the agreed field in the access request corresponding to the user access behavior, such as the protocol header of http or https or a certain data segment of the data unit; The network device 1 obtains the user-side access behavior, and extracts the user mark of the user in the corresponding appointment field of the access request corresponding to the user access behavior. Preferably, the network device 1 further includes a user identification library 106 to store user identification and other information, such as the correspondence between the device identification code and the user identification; the user identification library 106 is based on the correspondence between the device identification code and the user identification determined in step s7 The relationship updates the relevant content of the user ID library for further processing.

[0065] For those skilled in the art, it is obvious that the present invention is not limited to the details of the foregoing exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description, and therefore it is intended to fall within the claims. All changes within the meaning and scope of the equivalent elements of are included in the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved. In addition, it is obvious that the word "including" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claims can also be implemented by one unit or device through software or hardware. Words such as first and second are used to denote names, but do not denote any specific order.