146results about How to "Recommended results are accurate" patented technology

The invention discloses a group and place recommendation method based on locations and social relationships. The group and place recommendation method comprises the steps of: acquiring user check-in information in an LBSN, removing places and user data with poor effectiveness, and finally acquiring check-in data of users; utilizing a Pearson correlation coefficient, measuring check-in similarity based on common check-in data of the users, calculating a check-in similar degree among the users, and establishing a user check-in similar degree network; identifying different communities by utilizing a discrete particle swarm optimization method according to the user check-in similar degree network; and acquiring a friend list according to accounts of a user social network, forming social adjacent relationships of users in communities, finally generating a social group, recommending the social group to target users and recommending places to the target users by adopting the collaborative filtering recommendation method. The community finding method is simple and easy to operate, and the community dividing speed is fast. By adopting the method of combining the social group with place recommendation, the complexity of the method is reduced, and the recommendation precision is improved.

The invention discloses a combined wrong question recommendation method based on a knowledge graph. Wrong questions relevant to weak knowledge points of a learner can be accurately recommended for the learner by adopting the method. The method comprises the steps that knowledges are extracted from large-scale unstructured test question data to establish the knowledge graph; text mining and word segmentation are conducted on the wrong questions of the learner to extract wrong question keywords, and thus knowledge points including in the wrong questions are determined; semantic near neighbors of the knowledge points are obtained by analyzing semantic similarity of the test questions; the wrong question knowledge points are mapped into the knowledge graph to obtain test question entities conforming to their knowledge points. In addition, similarity weight calculation is conducted on a test question library to obtain similarity matrixes of test paper, a collaborative filtering technology is utilized to obtain recommended test questions of the wrong questions. Finally, two recommendation results are further combined in weighing, mixing, superposing and element-level modes, and a final recommendation result is given.

The invention discloses a real-time goods recommendation method based on customer shopping intention exploration. The method comprises the following steps: 1, constructing a tree-shaped customer shopping intention model; 2, detecting user behaviors in real time and identifying user intention; 3, detecting the change of the user intention in real time; 4, tracking the change track of the user intention and predicting the deep intention of the user: predicting the deep intention of the user based on the discrete intention of the user; and 5, determining a goods recommendation strategy: determining the goods recommendation strategy according to the shopping intention of the customer. The method disclosed by the invention automatically identifies the shopping intention of the customer, performs precise individualized recommendation, and improves the customer satisfaction; and moreover, the method automatically detects the change of the shopping intention of the customer and adjusts the goods recommendation strategy. The method adapts to the changing interest of the customer in real time, deeply explores the shopping intention of the customer based on the user intention model and provides professional recommendation service to the user.

The invention provides a recommendation method and a device of a text related subject. The recommendation method comprises the following steps: S1. acquiring content of a text, obtaining a term by carrying out the word segmentation on the content of the text, calculating the weight each term and confirming a head-word and an auxiliary word of the content of the text according to the weight of the term; S2. using the head-word to carry out a matching in an existing subject collection, using the subject which contains the head-word as a candidate subject to form a candidate subject collection; S3. calculating the comprehensive weight of each candidate subject according to corresponding resources of the candidate subject and the relativity between the candidate subject and the auxiliary word; and S4. using the candidate subject that the comprehensive weight meets preset requirements as a recommended related subject. Compared with the prior art, the recommendation method and the device of the text related subject can achieve the recommendation of extensive reading requirements of users, so that recommending results are more accurate and the recommendation method and the device of the text related subject is closer to the using habit of the users and meets the extensive reading requirements of the related subject.

The invention discloses a collaborative filtering recommendation method for integrating time contextual information, which is used for integrating the time contextual information on the basis of an original item-based collaborative filtering recommendation algorithm and an original user-based collaborative filtering recommendation algorithm and combining the original item-based collaborative filtering recommendation algorithm and the original user-based collaborative filtering recommendation algorithm into a uniform algorithm. The collaborative filtering recommendation method comprises the steps of for the user-based collaborative filtering recommendation algorithm, firstly, integrating a time attenuation function in a user similarity calculation stage; then, clustering items, and training interest attenuation factors of a user on an article category; finally, integrating the time attenuation function in a rating prediction stage, wherein for the item-based collaborative filtering recommendation algorithm, the process is similar to the process of the user-based collaborative filtering recommendation algorithm, and the two algorithms can be finally combined into the uniform algorithm. According to the collaborative filtering recommendation method disclosed by the invention, the time attenuation function is introduced in both the similarity computation stage and the rating prediction stage, different time attenuation factors are used for different types of items by different users, and thus the prediction accuracy can be effectively increased.

The invention discloses a personalized recommending method fused with user trust relationships and comment information. The method includes adding useful comments and user trust relationships on the basis of performing probability decomposition on a score matrix, wherein the user trust relationships explicitly show the trust level of users, the user useful comment information potentially shows the trust level of the users, and the two aspects of information can predict interests and hobbies of the users; and an alternating least squares is used to train model parameters. The personalized recommending method can fuse user trust relationships in a credible network and the potential trust relationships acquired by the useful comment behaviors, and can improve the recommendation precision.

A collaborative filtering recommendation method for integrating preference relationship and trust relationship includes the following steps: digging preference relationship between users according to a project grading datum of users and building a preference relationship network; integrating preference relationship and trust relationship and building a preference and trust relationship network; positioning similar neighbors of a target user in the Markov random walk method and on the basis of the preference and trust relationship network; and forecasting grading values of corresponding projects for the target user on the basis of the grading value of the similar neighbors on a certain project. By adopting the method, a recommendation system can forecast grades of the projects made by the users efficiently in a novel mode. Compared with the prior art, the collaborative filtering recommendation method has the advantages of being simple, easy to achieve and capable of generating accurate grading forecasting values; and being convenient to select due to the fact that the method has only one parameter and recommended results are insensitive to the parameter.

The invention provides a label based user collaborative filtering content recommendation method. The method includes: acquiring label information of a plurality of recommendation contents to be recommended; performing clustering on the label information according to similarities among the label information to obtain a plurality of label clusters; obtaining interest vectors of a plurality of users according to the label information and the label clusters; obtaining the similarities among all the users according to the interest vectors of all the users and first browsing histories of all the users, and determining users associated with a target user according to the similarities among all the users, wherein the first browsing histories include sets of contents which are browsed by the users, and browse time of the contents in the sets; and obtaining recommendations of all the contents to be recommended corresponding to the target user according to the similarities among the target user and the users associated with the target user and the first browsing histories of the users associated with the target user, and performing content recommendation on the target user according to the recommendations. The method is high in recommendation accuracy.

The invention discloses an algorithm for collaborative filtering recommendation based on interest changes and trust relations. The algorithm mainly comprises the steps that (1) users' interest similarity degrees are computed by a time-fusion attenuation function; (2) a trust network of the users are established, and the users' trust degrees are computed; (3) the users' similarity degrees are computed in combination of the users' interest similarity degrees and the users' trust degree; and (4) a score of a target user aiming at a project is predicted. According to the invention, based on computation of the users' interest similarity degrees with application of the time-fusion attenuation function, network modeling is conducted for evaluation relations between the users and the project, and trust relations between the users are analyzed and excavated. Finally, the user's interest and trust relations are synthesized in the collaborative filtering recommendation. In this way, recommendation accuracy is enhanced.

The invention provides a sequence recommendation method based on a dynamic interaction attention mechanism. The method comprises the steps of obtaining initial short-term preference and initial long-term preference of a user; obtaining long-term preferences and short-term preferences according to the interactive attention network in combination with the initial short-term preferences and the initial long-term preferences; and scoring the corresponding articles according to the sequence recommendation model in combination with the long-term preference and the short-term preference, and recommending the articles to the user according to the scoring result. According to the constructed dynamic interaction attention mechanism network model for sequential recommendation (DCN-SR), long-term and short-term interaction joint dependence expression of the user can be learned through the model, and the recommendation result is more accurate in combination with long-term preference and short-term preference.

The invention provides a personalized tour route recommendation method. The personalized tour route recommendation method comprises the steps that 1, travel note information of tourists is obtained and preprocessed; 2, a scenic spot type expression vector based on category information is obtained; 3, a touristy preference expression vector of each user, touristy preference expression vectors of the tourists in each month and an expression vector of each scenic spot are obtained; 4, a candidate tour route set is obtained according to step 1; 5, candidate tour routes are obtained from the candidate tour route set according to individual constraints; 6, a preference expression vector to which each tour route belongs is obtained; 7, similarity matching is carried out on the user touristy preference and the candidate routes to obtain the playing route which is most matched with the user touristy preference to serve as a final tour route recommended to the users. In the method, according toeach scenic spot in a historical tour track of the tourists and the scenic spot category information to which the scenic spots belong, the personalized touristy preference of the tourists is obtained.

The invention provides a method and a device for search recommendation based on a search term and a search engine. The method comprises the steps that multiple candidate recommendation objects are acquired; the similarity between every two candidate recommendation objects is acquired; the candidate recommendation objects are classified according to the similarity to generate multiple clusters; and the search term input by a user is received, and the candidate recommendation object relevant to the search term is acquired from the clusters to serve as a recommendation result. The technical scheme overcomes the defect that in the prior art, a recommendation result is excessively generalized; the clusters are utilized to divide the recommendation result, each cluster is accurately described through a knowledge map, valuable information is provided, the selected recommendation result better conforms to the search habit and search demand of the user, the recommendation result is more accurate and more professional, and the satisfaction of the user is improved.

The invention provides a personalized recommendation method and system for friends and applications. The method comprises the steps that A, a recommendation score S is calculated; B, recommended content is displayed according to the recommendation score; C, a recommendation weight vector is adjusted according to feedback information about the recommended content from a user; D, the recommended content is updated according to the adjusted user recommendation weight vector. The personalized recommendation method and system have the advantages that the application range is wider, recommendation results are more accurate, the calculating speed is higher, calculation at idle time is achieved, a blacklist mechanism is adopted, the accuracy of information recommendation can be effectively improved, calculation cost is saved, and a good recommendation effect is achieved.

The invention provides a highly accurate code snippet recommendation method. The method comprises the steps of splitting and dividing a project, and generating class files, a code snippet , a code snippet method body and a code snippet note; generating a code snippet method body vector and a code snippet note vector based on the code snippet method body and the code snippet note; compiling and decompiling each class file, and extracting an instruction sequence of the code snippet from the decompiled file; generating an inquiry vector based on an inquiry input by a user, and screening out a group of primary recommendation results with highest text similarity according to a text similarity of the inquiry vector and a code snippet text characteristic; and according to statement similarity ofall code snippet s in the primary recommendation results, screening the primary recommendation results for the second time and then re-ranking so as to form a final recommendation result. The method overcomes the defect that the existing method is low in recommendation accuracy due to unique use characteristic and failure in filtering of recommendation results. The method is applicable to the field of recommendation of open source code snippets.

The invention relates to television program recommendation. The invention provides a method and a system for recommending digital television program based on bayesian algorithm aiming that existing television program recommendation method only statistics user's viewing behavior but ignores playing behavior of television station. The method includes: collecting user's viewing recorder; analysing user's viewing recorder and obtaining user's viewing action parameter. The method also includes: obtaining television station playing action parameter according with viewing behavior parameter; predicting user's viewing probability to waiting-played television program according with viewing action parameter and playing action, and recommending television program to user thereof. The invention also provides a recommendation system corresponding to the method. Because combining the user's viewing action parameter and television playing action parameter when predicting user viewing special program, a recommendation result obtained by the television program recommendation is more accurately and reflects real user's viewing taste.

The invention discloses a scenic region route recommending method based on sightseeing behaviors, and a system of the same. The scenic region route recommending method based on sightseeing behaviors includes the steps: acquiring the behavior data, generated from the sightseeing process, of a tourist, and extracting and generating a group of candidate sightseeing routes from the acquired historicalsightseeing behavior data by means of a frequent sightseeing route extraction method; and according to the total sightseeing time, the sightseeing starting / terminate place, and other personal sightseeing constraints input by the tourist, searching the route which can satisfy personal constraints and has the highness sightseeing value from the candidate routes and recommending the route to the tourist. The scenic region route recommending method based on sightseeing behaviors has the advantages of high accuracy of the recommending route, high customized degree of the recommending result, and high sightseeing experience degree.

The invention discloses a cross-domain recommendation method fusing label and attention mechanism and an implementation system thereof, and the recommendation method comprises the steps: firstly, selecting and constructing a cross-domain fused label, and respectively carrying out the weighted summation of label vectors of a source domain and a target domain to obtain a resource vector; secondly, according to an interest mining algorithm based on an attention mechanism, obtaining preferences of the user in a source field and a target field; thirdly, learning label mapping between a source domain and a target domain according to a BP neural network-based cross-domain label mapping algorithm, and obtaining comprehensive preferences of the user in the target domain; and finally, recommending projects with high similarity to the comprehensive preferences of the users in the target domain to the users through a cross-domain recommendation algorithm fusing label mapping and attention mechanisms. The preferences of the user in different fields are comprehensively considered through cross-field recommendation, so that the cold start problem of the user in target field recommendation is improved; and meanwhile, in a cross-domain recommendation system, by analyzing the preferences of the user in different domains, the recommendation results are more diversified.

The invention relates to a personalized travel package recommendation method based on demand classification and subject analysis. The method includes: analyzing natural language form demands input by a user, utilizing word segmentation, demand classification and other natural language processing techniques to process and classify the user demands so as to obtain rigid demand, flexible demand and negative demand of the user; then utilizing an LDA (latent dirichlet allocation) document theme generation model, making travel service individuals effectively cluster into different service fields by theme similarity, and then conducting similarity matching with the user demands so as to obtain a service list best matching user expectation; finally carrying out travel package design recommendation by means of travel package optimized recommendation algorithm: firstly acquiring a travel package scenic spot set according to user time demand and service priority information; then combining location information, preference information and the like to determine travel package hotel service; and then calculating the optimal journey of every day according to the distance function L, and ranking the travel package according to travel package recommendation index; and finally, selecting travel package catering service according to scenic spot location and user preference. By integrating the processing, the purpose of designing and recommending personalized travel package best meeting the user demand can be realized.

The invention provides a method for solving collaborative filtering recommendation data sparsity based on a neural network. The method for solving collaborative filtering recommendation data sparsity based on the neural network adopts generalized regression of neural network (GRNN) and conducts full filling on sparse data by a train network model and score prediction. The method for solving collaborative filtering recommendation data sparsity based on the neural network comprises the following steps: before conducting the GRNN training, conducting screening on input variables of the neural network by adopting mean impact value (MIV), choosing characteristic values having great impact on output as valid input variables; using the valid input variable to construct the input matrix of the GRNN; adopting Kfold cross validation circulation to find out an optimal spread value of the GRNN; using the optimal spread value and the corresponding input matrix and output matrix to conduct GRNN training; using the trained GRNN to conduct score prediction on a sparse score matrix; and replacing non-scored data of the sparse score matrix with predicted score values. The method for solving collaborative filtering recommendation data sparsity based on the neural network can conduct fully filling on sparse recommendation data, solve the data height sparsity problem most outstanding in existing collaborative technology, and enable recommendation result to be accurate.

The invention discloses a social network user recommendation method based on the extraction of the user interest and the social topic. The method comprises the following steps: obtaining the ''follower-followee'' relationship and the ''mutual-following'' relationship information of a user in a social network; extracting the social topic and the interest topic of the user; on the basis of the topics, independently forming the social community and the interest community of the user; and independently tidying each community into a matrix form, using a matrix decomposition method to decompose each community matrix to calculate the intention following score of each user on each community matrix, taking the maximum value of the intention following scores of each user on all communities as a final intention following score, sorting the final intention following score between a target user and other users for the target user, and finally selecting the TOP-N users with the highest score as recommended following users. The method has the advantages that an accurate social network user recommendation result is obtained.

The invention discloses an ability prediction and recommendation method and device of a crowd-sourcing software developer. The method comprises steps of determining the task difficulty level and the task score of a developed task of a developer, and according to the task difficulty level and the task score, calculating a development ability value for the task of the developer; classifying historytasks to obtain multiple task categories, wherein each task category comprises more than one task; based on each task category and the development ability value to each task of each developer, calculating development ability curves in different task categories of each developer; by use of negative exponent type learning curve model, fitting the development ability curves in different task categories of each developer so as to obtain learning curve models under different task categories of each developer; and based on the learning curve models under different task categories of each developer,recommending a developer to a target task.

The invention relates to a hidden feedback recommending method of a multiple-GRU-layer neural network based on a user space and a system thereof. The method comprises the steps of through mapping an object to the user space, selecting an attention object characteristic for a user, performing multidirectional time sequence analysis according to a BP or PBTT algorithm and a recursion structure whichis unique in a multiple-GRU-layer neural network, predicating an interested object of the user according to the object sequence which is accessed by the user, and supplying a recommended service forthe user. The system comprises a historical behavior collecting module for an object which is accessed by the user, an object mapping user space module, a multiple-GRU-layer neural network training module and a recommending list generating module. Based on the system of the invention, the user behavior with relatively high randomness can be more accurately represented, and more accurate recommending result is obtained.

The embodiment of the invention discloses a method and apparatus for generating a recommendation result. The recommendation result can be more accurate. The method disclosed by the embodiment of the invention comprises the following steps: obtaining behavior records of a user on at least two application platforms; generating user features according to the behavior records, and obtaining attribute features of a business of a target application platform; generating joint features according to the user features and the attribute features; generating a sample set according to the joint features; obtaining a logistic regression model according to the sample set; calculating the probability of a to-be-recommended business according to the logic regression model, wherein the to-be-recommended business is the to-be-recommended business on the target application platform; and generating the recommendation result according to the size of the probability of to-be-recommended business.

The invention provides a short video recommendation model based on short video multi-modal features, and the model comprises the steps: 1, carrying out the feature extraction of short video title features through employing a TF-IDF method, and reducing the dimension of a short video title feature vector to k dimension through employing a PCA dimension reduction algorithm; 2, extracting 128-dimensional original features of the short video content, and adopting a PCA dimension reduction algorithm to reduce the dimension of the feature vector of the short video content to k dimension; 3, extracting 128-dimensional original features of the short video background music, and reducing the dimension of the feature vector of the short video background music to k dimensions through a PCA dimension reduction algorithm. According to the invention, the influence effects of the feature data of different modes on the user behaviors generated by the user are considered to be different; and learning the influence proportion of different modal data of the short video on the user by using a hidden Markov model, and mapping the multi-modal features of the short video to a unified vector space for fusion based on the influence proportion to obtain short video feature data represented by the multi-modal data features.

The invention provides a scholar recommendation and cooperative prediction method based on representation learning and competition theory, belonging to the field of computer software. With the huge academic network, the recommendation needs are met, through the presentation learning technology, the paper published by the author and the network relationship among scholars are analyzed, with the help of competition theory, the time conflict is solved in order to obtain the recommendation results and collaborative forecasting. The method obtains valid data from Microsoft dataset and preprocessesthem, constructs a dynamic model for calculating the similarity of scholars' personality based on the content of the paper, constructs a collaborative network-based model for computing the similarityof academic environments, and constructs the processing model of competition theory, the model is trained with the preprocessed data set, according to the personality similarity obtained from the training, a preliminary recommendation list is generated to weaken the source and target scholars who are too similar by using the environmental similarity, and the time conflict is eliminated by using the processing model of competition theory, so as to achieve effective scholar recommendation and the next collaborative object prediction.

The invention relates to a smart television technology and provides a personalization recommendation method based on a smart television. The problem that an existing smart television recommendation method is single in recommendation technology is solved. According to the technical scheme, a system firstly collects user personal characteristic information and user personal preference data, performs modeling according to the user personal characteristic information and the user personal preference data, generates a user similarity model, then analyzes the correlation of all programs in film database, conducts program similarity modeling through content filtering to generate a program similarity model, grades the programs in the program similarity model through the user similarity model to obtain a user-program grading matrix, and adopts a collaborative filtering algorithm to filer the user-program grading matrix so as to obtain a recommendation result and returning the recommendation result to a user when the recommendation result is needed. The personalization recommendation method based on the smart television has the advantages of bringing convenience to the user and being suitable for a smart television system.

The embodiment of the invention discloses an information recommendation method and device based on historical data records, equipment and a medium, and relates to the technical field of computer information processing. The method comprises the steps of receiving an information recommendation instruction; identifying a target group to which the target user belongs; calling historical browsing records of the target user and the target group, and intercepting a first historical record and a second historical record from the historical browsing records; obtaining a target browsing item; calculating a first recommendation score s1 and a second recommendation score s2 of each target browsing item, and obtaining a first weight w1 and a second weight w2; calculating a recommendation score s corresponding to each target browsing item; and generating a user recommendation list and sending to the target user. The invention also relates to a blockchain technology, the recommendation score is stored in the blockchain network, the method can effectively avoid the problem of poor recommendation effect when the personal browsing habit of the user is not concentrated, the recommendation result is more accurate, and the recommended content can better meet the interest demand of the user.

The invention discloses an employee individualized-learning recommendation method based on a learning map and collaborative filtering.The method includes the steps that resource features and employee attributes are respectively extracted according to the learning resource content of an online learning platform and the practical conditions of learners (enterprise employees), a mathematical model is built, a recommendation list is calculated according to similarity to generate the recommendation result, the feedback conditions of the learners are collected to be used for improving similarity calculation, and the recommendation process is optimized.The method has certain universality in recommending the content of semi-structured data, unstructured data and multimedia learning resources, the learning map of the employees and collaborative filtering are combined, the recommendation effect is corrected and optimized, the sparse scoring matrix and learning resource recommendation, namely, cold starting, of new employees can be effectively achieved, pushing of the learning content of the online learning platform is more user-friendly, the enterprise employees are effectively assisted in rapid growing, employee training and learning cost is saved, and employee learning efficiency is improved.

The invention relates to a catering resource recommendation method based on a video application and a system thereof and belongs to the information pushing field. A problem that routine video application pushing information does not match with a user demand so that a user experience is poor is solved. The method is characterized by acquiring characteristic information of a user when the user operates the video application, and according to the characteristic information, acquiring a plurality of characteristic labels of the user; matching the plurality of characteristic labels of the user with a plurality of preset characteristic labels of catering resources in a catering resource database one by one so as to acquire a matching degree of the user and the catering resources in the catering resource database; and according to the matching degree of the user and the catering resources in the catering resource database, screening the catering resources suitable for the user and pushing to the user. The method and the system are used for accurately recommending catering resource information to the user in a customized mode, an experience of the user to use the video application is increased and maneuverability of the video application is enhanced.