39results about How to "In line with interest" patented technology

The present invention provides an advertisement click rate estimation method based on a decision tree and a device. The method comprises a step of obtaining the related characteristic information of a specific historical advertisement in a predetermined historical time period, a step of obtaining the personalized characteristic information of a target user, a step of carrying out decision tree model calculation based on the obtained target user personalized characteristic information and the related characteristic information of the specific historical advertisement so as to determine a cross characteristic vector for estimating the click rate of the specific historical advertisement, wherein each leaf node of the decision tree model represents one cross characteristic, a step of calculating based on the cross characteristic vector and the model training parameter obtained from training in advance so as to estimate the click rate of the specific historical advertisement. The invention also provides an application software recommendation method and a device. According to the method or the device, through the decision tree model, the cross characteristic with a strong classification characteristic is generated, and thus the advertisement click rate is accurately and rapidly estimated.

The invention provides a child companion robot and a child companioning system. The child companion robot comprises a speech input module, a touch input module, a voice output module and a display module. The speech input module extracts speech information of a user from surrounding voices of the child companion robot; the touch input module receives touch operations of the user through an interactive interface; the voice output module outputs response speeches responding to the speech information or feedback voices responding to the touch operations; the display module outputs the interactive interface or expression images matched with the speech information and the response speeches. The child companion robot and the child companioning system have the advantages that various operations of the user can be responded intelligently to provide good user experience, a companioning function of the child companion robot is achieved, and energy and time of patients for companioning children are saved.

The embodiment of the application discloses methods, devices and equipment of information recommendation. One of the methods includes: acquiring page data of a first link in a first page and related links or data of the first link when clicking on the first link is detected; and displaying the page data of the first link, and loading the related links or data of the first link in the first page. According to the embodiment of the application, recommendation information (namely the related links or data of the first link) related to the page data of the first link is obtained through currentlyclicking on the first link by a user and is recommended to the user, The information recommendation is enabled to not need to completely rely on currently and commonly used personalized recommendationtechnology, recommended information is the information related to the page data currently viewed by the user, thus the recommended information is enabled to more accord with current interest of the user, probability of clicking on and viewing the recommended information by the user is increased, and then a success rate of information recommendation is increased.

The invention discloses a dynamic user attribute extraction method based on social media. The dynamic user attribute extraction method includes the steps: firstly, performing text preprocessing for an acquired training sample set and extracting subject terms to obtain K subjects and m subject terms of each subject; secondly, extracting short texts of a user to be processed, dividing time sub-segments, filling data through a time sliding window to obtain text data of the time sub-segments, counting the occurrence frequency of the subject terms after text preprocessing to obtain attribute weight information of the subjects, introducing time attenuation coefficient, sequentially acquiring user attribute features associated with time attributes in time sequence, extracting the user attribute features of the latest time sub-segment as current user attribute features and outputting the current user attribute features. Without external knowledge, the short texts of the social media are semantically expanded by disordered words in the texts, and dynamic user attributes can be extracted from micro-blog texts released or forwarded by users.

The invention relates to a graph structure-based recommendation method, apparatus and device, and a storage medium. The graph structure-based recommendation method comprises the steps of obtaining attribute information of each video; According to the attribute information, calculating the common similarity between every two videos; Generating a candidate recommendation set according to the generalsimilarity; Obtaining a user historical operation record of the user identifier according to the user identifier; Calculating a recommendation weight parameter according to the historical operation record of the user; Calculating the final similarity between each video and the user identifier according to the recommendation weight parameter; And sending a video recommendation result to the user identifier according to the final similarity. The video recommendation method has the effects of improving the video recommendation result and enabling the video recommendation result to be closer to the interest of the user.

The invention discloses an interest resource recommendation method based on multi-dimensional attribute attention. A resource recommendation terminal which is in communication connection with a network system with resources is arranged; the resource recommendation terminal is provided with a visual device, a statistic module, a user type dividing module, an attribute attention guide module, a resource grading module, a topN choosing module and an attribute relevancy guide module. By the recommendation method, corresponding attribute attentions can be calculated according to different users, and resources which are most similar to user interests can be recommended. Excellent interest resources can be recommended even though a user does not browse the resources, and when the user browses the resources, similar predicted interest resources which are relevant to the resources are recommended.

The invention provides a method and device for recommending search terms based on input search terms. The method comprises the following steps: determining a matching degree between a search term input by a user and an application in a preset application library; using a label set belonging to the application having the highest matching degree as a label set of the input search term; determining first similarity between every two applications in the preset application library; determining second similarity between the input search term and other search terms in a preset search term set based on the matching degree, the label set of the input search term, and the first similarity; and selecting a certain number of search terms from the preset search term set in a predetermined manner based on the second similarity to be recommended to the user.

The invention provides a television video recommendation method and device, a server and a storage medium, and relates to the technical field of digital televisions. The method comprises the followingsteps: receiving an operation request sent by television equipment, wherein the operation request comprises identification information of the television equipment and user behavior data for operatingthe television equipment; according to the identification information, obtaining a video matrix corresponding to the television equipment, a video recommendation model corresponding to the televisionequipment and family portrait features of the television equipment, wherein the video matrix comprises recommended videos of multiple roles; adopting a video recommendation model to process the userbehavior data, the family portrait features and the video matrix so as to determine a recommended video of the target role from the video matrix; and pushing the recommended video of the target role to the television equipment. According to the invention, personalized recommendation service can be provided for each family role, and the recommendation effect is improved.

The invention discloses an information recommendation method and device and a computer readable storage medium, and relates to the technical field of the Internet, and the method comprises the steps:generating a first information recommendation list corresponding to a current user based on the information related to the historical behavior data of the current user and a pre-constructed first similarity matrix, and each similarity in the first similarity matrix represents the content similarity degree between every two pieces of information; forming a second information recommendation list corresponding to the current user based on a pre-constructed user information preference matrix and a second similarity matrix, and each similarity in the second similarity matrix is obtained based on the number of common preference users owned by every two pieces of information; and fusing the first information recommendation list and the second information recommendation list to obtain a final information recommendation list so as to recommend the final information recommendation list to the current user. The embodiment of the invention can improve the information recommendation effect.

The invention relates to the field of article recommendation, discloses a recommendation method based on a self-attention mechanism, and aims to improve the training efficiency of a recommendation model and the effect of personalized recommendation. The method comprises the following steps: firstly, collecting and preprocessing historical interaction information of a user to form a training sample set; designing a recommendation model, taking the training sample set as the input of the recommendation model, and adopting a square loss function as an optimization target to train the recommendation model; finally, calculating through the trained recommendation model to obtain the interaction probability of the user and to-be-recommended items, sorting the interaction probability according to the size, and generating a recommendation candidate set of the user.

The invention discloses an improved financial product recommending method by a content-based linear regression mode. When a product is recommended by using the existing content-based linear method, some coefficients in the generated regression equation are negative ones and thus the characteristics are in negative correlation with product purchasing by the user,; and the negative correlation is inconsistent with the actual application scene and an overfitting phenomenon is generated. With the method disclosed by the invention, a factor causing an error is found out directly and thus improvement is carried out, so that optimized recommendation meeting the practical situation is provided.

The invention discloses a news recommendation method based on multi-feature fusion, and the method comprises the steps: obtaining a multi-feature vector of each piece of to-be-recommended news in a to-be-recommended news set when a triggering condition for news recommendation to a target user is reached; determining a user portrait of the target user; determining a first recommendation result based on the user portrait and the topic feature vector of each piece of to-be-recommended news in the to-be-recommended news set; determining a second recommendation result based on the currently browsednews of the target user and the keyword vector, the title and the abstract feature vector of each piece of to-be-recommended news in the to-be-recommended news set; and determining a final recommendation result according to the first recommendation result and the second recommendation result, and recommending the final recommendation result to the target user. By applying the technical scheme provided by the invention, the news recommended to the user more conforms to the interest of the user and is more targeted, and the browsing experience of the user is improved. The invention further discloses a news recommendation device based on multi-feature fusion. The news recommendation device has corresponding technical effects.

The invention discloses a method and device for constructing and using a knowledge model of a session system, and the method for constructing the knowledge model of the session system comprises the steps: abstracting at least one theme associated with a knowledge point from the knowledge point in the session system; constructing a system theme tree based on the at least one theme; determining an association relationship among the knowledge points based on the system topic tree; and constructing a session system knowledge model at least based on the association relationship between the system topic tree and each knowledge point. Knowledge points are abstracted into at least one topic.. A topic tree is constructed by utilizing a topic, the association relationship among the knowledge pointsis determined according to the topic tree, and then the knowledge model of the session system can be constructed for recommending questions or answers to the user in the subsequent heuristic session process, so that questions or contents which more accord with the interest of the user can be recommended to the user.

The invention relates to an object recommendation method and device, electronic equipment and a storage medium, and relates to the technical field of information dissemination. The method comprises the steps: receiving an access request sent by an account on terminal equipment, wherein the access request carries account information, and the number of historical interaction operation information corresponding to the account information is smaller than a preset threshold value; determining a current state of the account based on the account information, and screening out a recommendation objectmatched with the current state from an information candidate library; and sending the recommendation object to terminal equipment, and receiving operation information returned by the terminal equipment, the operation information being used for updating the current state, so that object recommendation is continuously performed according to the latest returned operation information in the process ofinteracting with the user, and the recommended recommendation object more accurately conforms to the interest of a new user.

A movie recommendation system based on an improved deep structured semantic model comprises a user behavior collecting and processing module, an offline training module and an online recall and sorting module, the offline training module receives merged data output by the user behavior collecting and processing module, and the online recall and sorting module takes out user feature vectors obtained by training of users according to attribute features of the users and recall movie subsets recommended for the users in a movie vector library by adopting an approximate nearest neighbor search technology; a movie recommendation method based on an improved deep structured semantic model comprises the steps of user behavior collection and processing, off-line training, online recall and sequencing and the like, movies conforming to user interests can be effectively mined according to dominant features and implicit interaction information of users and movies, personalized recommendation services are provided for the users, and the movie recommendation efficiency is improved. And obtaining a recommendation result of the user.

The invention provides a live broadcast room recommendation method, a content recommendation method, a terminal and a storage medium. The live broadcast room recommendation method comprises the following steps: obtaining an interaction record of a target user about a product and/or a live broadcast room; live broadcast data corresponding to current live broadcast of each candidate live broadcast room is acquired, and the candidate live broadcast rooms are live broadcast rooms for live broadcast of at least one product; and according to the interaction record, the live broadcast data and a pre-trained recommendation model, determining a matching degree between each candidate live broadcast room and the target user, recommending at least a part of the candidate live broadcast rooms to the target user according to the matching degree, and through the model in combination with an interaction behavior of the user and a live broadcast behavior of the live broadcast room, recommending the target user to the target user. Therefore, the live broadcast room recommendation accuracy is improved, and the live broadcast watching experience of the user is improved.

The invention aims to provide a method and an apparatus used for providing to-be-presented information in a mobile device. The method comprises the following steps that an operation system sends a triggering signal to at least one corresponding application to start the at least one application when it is detected that the mobile device is in a specific region. The method and the apparatus have the advantages that the mobile device is detected by utilizing a module contained in the operation system, and the starting of the specific application is triggered when it is detected that the mobile device is in the specific region, so that normal operation of application functions still can be ensured when certain applications are not started, the waste of resources of the operation system due to continuous operation of the applications is reduced, and the usage experience of users is improved.