109 results about "Recommendation quality" patented technology

The invention discloses a video recommendation method and system. The method comprises the steps that an initial score of a user on a video is calculated according to the length of time of the video and the length of time spent by the user in watching the video; the initial score is standardized in a minimum-maximum standardization mode to obtain a standardization score of the user on the video; according to the difference value between a current moment and the moment when the user watches the video last time, calculating a final score of the user on the video. According to the video recommendation method and system, the influences of the playing time of the video, the user watching frequency and time interval factors on user interestingness are comprehensively considered, and the final recommendation quality is improved.

The invention relates to an individual behavior and group interest-based interest place recommendation method and device, and belongs to the field of place recommendation. The method comprises the following steps: determining which category of places users are about to go in the future through analyzing the individual behaviors of the users; dividing the users into groups through extracting features of the users; and sorting the places, belonging to the place category determined in the first step, in the groups of the users, so as to recommend specific places for the users. According to the method and device, the individual behaviors of the users, and the distance relationship between the interests of the groups of the users and the sign places of the users are comprehensively considered, and the place category is used as a bridge for connecting the individual behaviors and the group interests, so that the user interests are mined and recommendation is carried out. Compared with the prior art, the method and device are not only capable of improving the recommendation quality, but also capable of greatly reducing the time complexity of a recommendation algorithm so as to improve the recommendation efficiency.

The embodiment of the invention provides a video recommendation method and device. The method comprises the following steps of: receiving and displaying one or more video recommendation data to enable a user to select a target video according to the video recommendation data; displaying the target video in a preset display area when an event that the user triggers the target video is detected; and monitoring the operation of the user on the display area, and sending a monitored operation result to a server to enable the server to update the video recommendation data according to the operation result. Through adoption of the method, the video recommendation quality and the user experience are improved.

The invention discloses a commodity recommendation method based on group community. The method comprises the following steps: analyzing historical shopping behaviors of a user, extracting shopping operation information and establishing a user interest model; carrying out clustering analysis based on user interests on the extracted shopping operation information to enable the users having the sameinterests to be grouped into one community; analyzing shopping behavior history to explore interest circles with similar interests, and carrying out on-line commodity recommendation; and based on theuse shopping history, carrying out reverse commodity category analysis, and carrying out commodity recommendation according to the category of the commodities. The method realizes layer-by-layer deepanalysis from three layers, that is, the user layer, the commodity classification layer and the commodity category layer, so that a recommendation system can identify and recommend the favorite commodities of the users more and more accurately according to similarity features between the users, calculation amount is reduced greatly, and the problem of bottleneck of online recommendation speed canbe relieved effectively; and through the identified commodity community, associate shopping information can be identified more effectively, and recommendation quality is improved.

The invention discloses a failure record (FR) method for optimizing personalized recommendation results, which improves the personalized recommendation quality and precision by using social tag network filter and recommendation deviation removal. The social tag network filter method comprises the following steps of: establishing a project social network K neighbor by using a social tag network model, and constructing a social tag filter set during recommending in a recommendation model based on the project social network K neighbor, wherein the social tag filter set is used for filtering recommended projects with low social tag relevance in the user scored projects in the recommendation results of a project-orientated K neighbor model so as to combine information in user-project scoring data and social tag network data to recommend. The recommendation deviation removal comprises the following steps of: based on prediction values of the project-orientated K neighbor model on the known user-project scoring data and a turn score of the user, estimating the recommendation deviation by using a linear model; and when the recommendation is performed by using the recommendation model, removing the corresponding recommendation deviation estimation values from the scoring prediction values so as to optimize the recommendation results.

The invention provides a method for establishing a MicroBlog user interest inference model. The method comprises an interest label calculation model, an interest model used for MicroBlog text content extraction and a blogger interest point model used for blogger social relationship extraction, and the three models are fused through a model fusion strategy to obtain the final MicroBlog user interest inference model. The method combines personal information, MicroBlog contents and the social relationship, adopts a USER strategy that all MicroBlog contents of the same blogger are mixed by aiming at the sparsity problem of the MicroBlog contents, mines an implicit theme of the MicroBlog by a LPA (Label propagation algorithm), puts forwards a social label propagation algorithm on the basis of a network formed by blogger attention, and calculates influence on the blogger by various interest labels. The method exhibits good identification capability and information filtering capability, and filters false information to identify false bloggers before recommendation is carried out, so that the recommendation quality and accuracy of a recommendation system can be improved, and better experience is brought for the blogger.

The invention provides an information recommendation method and system. The method includes processing users, items and grades of the items and from the users into a bipartite graph; determining the similarity between a target user and other users according to the grades of the items and from the users; utilizing vector composed of the similarity of the target user and other users as a restarting vector and calculating relevancy between the target user and the items based on a restartable random walk model; selecting information that the items correspond to and recommending the information to the target user according to the relevancy. According to the information recommendation method, the relevancy between the target user and other users is determined according to the grades of the items and from the users, the relevancy between user nodes and item nodes is completely excavated, the data sparse problem effect is greatly reduced, the grade items that the users are interested in can be accurately recommended, and the recommendation quality is improved.

The invention discloses a collaborative filtering recommendation method based on association rule prediction, and relates to the field of personalized recommendation of the Internet. According to the feedback of searching and browsing behaviors of users, item coring data by the users are mined and obtained; the items which can not be mined are predicted by mining association rules between user characteristics and item characteristics; a scoring preference matrix of the users is constructed through the association rules between the users and the items; according to the preference matrix of the users, a sparse scoring preference matrix of the users is predicted; a traditional similarity measuring method is improved, the similarity between the user ui and the user uj is calculated, the most adjacent user BNS of a target user is obtained through the user similarity method, and recommended scores are obtained finally through the BNS; the collaborative filtering recommendation method based on association rule prediction is evaluated. According to the method, the accuracy for calculating the similarity between the users is effectively improved, high recommending quality is kept under the condition of sparse data, and the recommending quality for new users is also high in readiness.

The invention provides a user similarity-based sparse data collaborative filtering recommendation method and mainly aims to solves the problems in the prior art that the calculation of the values of similarity between users is in accurate for sparse data and further the recommendation quality is influenced. The method comprises the steps of (1) establishing a sparse matrix for item scores from users; (2) calculating the overall situation similarity between any two items; (3) calculating the local similarity between any two user scores; (4) calculating the similarity between any two users; (5) predicting scores; (6) generating a recommendation list; (7) completing collaborative filtering item recommendation for all the users. Experimental simulation results show that for sparse data sets, the method has the advantages of guaranteeing higher accuracy of the similarity between users, improving the recommendation quality and better meeting user requirements compared with conventional collaborative filtering recommendation methods.

The invention provides a Storm-based real-time recommendation method and a system therefor. The real-time recommendation method comprises the steps of: A, performing offline similarity calculation on video data and behavior data of users separately to obtain a similarity model of video, and performing statistical analysis on the behavior data to obtain the statistical data of user behavior; B, performing user modeling according to the video data in the Storm end to generate interest vectors and combining the interest vectors and the similarity model to obtain individualized recommendation results based on the user behavior; C, integrating the statistical data, the similarity model and the individualized recommendation results to obtain a recommendation list. The method combines similarity and interest, thereby improving the accuracy of individualized analysis of user preferences and guaranteeing the recommendation quality.

The invention relates to an internet subject ranking system and a method based on user prestige. The invention relates to the internet field and solves the problems of too simple basis, evenness and subjectivity when various person subjects, affair subjects and objects subjects on the internet are ranked. The invention introduces the concept of user prestige during the ranking process. A user input subject information for recommendation; different users grade the recommended subject; the system calculates the score and the ranking of the subject according to grades and the prestige of the users participating in the evaluation and calculates the user prestige of the presenter according to the score of the recommended works, and finally integrates the subject ranking and related information and then reveals the subject ranking and the related information to all the users. By adoption of the system and the method, correlation and interaction of recommendation and evaluation are realized; the ranking system can gradually distinguish the users into valuators with different appreciation abilities and endow the users with different evaluation influences through analyzing recommendation quality of the users, thereby the ranking results are more objective and more persuasive.

The invention discloses an individual recommendation method and a device based on fusion strategy. The individual recommendation method comprises the following steps: respectively confirming a first score value obtained by using a collaborative filtering recommendation method based on users, of a target user on a target item, confirming a second score value obtained by using a recommendation method based on image similarity, of the target user on the target item, and subsequently adding the weights of the first score value and the second score value so as to confirm a final score value of the target user on the target item, that is, the score values predicted by using the two methods are comprehensively considered. Therefore, the problems that a single recommendation algorithm is low in recommendation quality and the reliability is low are solved.

The invention relates to a random walk and diversity graph sorting-based personalized service recommendation method. Firstly the defect in calculating user similarity by a conventional Pearson correlation coefficient is analyzed, and then a similarity relationship between users is subjected to multi-time transmission through a WRW model, so that more similar neighbors are found for a target user,and the data sparsity problem is effectively solved; based on predicted QoS values of all the similar neighbors, a service graph model is built by adopting an SGMC algorithm to filter a large amount of candidate services with excessively low performance, reduce an algorithm optimization space, and ensure quick convergence of a subsequent diversity graph sorting algorithm; and finally, an optimal node set is searched for on the service graph model by adopting an ONCS policy, and k nodes contained in the set are a service list recommended to the user and having recommendation accuracy and functional diversity. The method has relatively high recommendation accuracy and relatively good functional diversity, and can meet potential functional demands of the users to the maximum extent while ensuring service recommendation quality.

The invention is a kind of synergistic filter recommending method for processing super-large user quantity, which belongs to network technology field. The method is: the point data of the project by the user are memorized in distributed way, namely that each user stores the point data of itself project, and acquires the point data of similar user through the distributed hash table, and acquires the predicted value through the synergistic filter method, then acquires the recommended value further, and restrains the quantity of similar user returned from the network covered by the hash table of each project, and adjusts the affection of each user, makes the affection of user which has more similar number and point two-tuples is larger, thus it upgrades the accuracy of prediction. The invention introduces the hash table route algorithm into the synergistic filter system, and it is improved, it solves the problem that the extensible ability of current filter system is bad and it upgrades the recommendation quality.

The invention provides a context-perceptive IPTV program recommending method. The method comprises the steps of calculating a concealed score of a watched program from a user, a confidence coefficient and context weight according to a watching record of a user; regarding each user and each program, initializing a user vector and a program vector according to a context; carrying out dimensionality reduction on the user vector and the program vector; utilizing a semanteme model to carry out scoring prediction to form recommendation. The context-perceptive IPTV program recommending method improves a traditional IPTV program recommending strategy according to the context information and has the advantages of analyzing user behavior according to the context, conforming with the practical situation better and having higher prediction and recommendation quality.

The invention provides an information processing method and a device thereof. The method comprises the following steps: acquiring a project score dataset of users, wherein the project score dataset comprises scores on the projects of users; acquiring a first relevance weight Wi between each node and a user node to be recommended in a user-project score database, wherein the first relevance weight serves as a relevance weight in the same type of projects; calculating a second relevance weight We between each node and the user node to be recommended in the user-project score database, wherein the second relevance weight serves as a relevance weight between different types of projects; acquiring the recommendation degree of each node according to We and Wi; and determining projects to be recommended according to the recommendation degree. By virtue of the information processing method and the device thereof, Wi is firstly calculated by using a random walk mode based on the characteristics of a bipartite graph, and the users are implicitly classified; secondly, the relevance between the users to be recommended and the neighborhood users is directly calculated and serves as We in order to avoid the repeated calculation of the relevance between the nodes, so that the potential preferable project nodes of the users are fully exploited, and the recommendation quality is improved.

The invention discloses a movie marking prediction method based on implicit bias and interest of friends. The method includes the steps of constructing a marking prediction model integrating personal bias with the interest of the friends according to movie marking logs of users and friend list data of the users, conducting model training to obtain bias matrixes of the users on movies, user implicit feature matrixes and article implicit feature matrixes, generating a friend interest implicit feature matrix through combination with the friend list data and the user implicit feature matrixes, generating a user-article prediction marking matrix according to the user implicit feature matrixes, the article feature matrixes, user-article bias matrixes and the friend interest implicit feature matrix, and finally generating a recommendation list for the users. Through the movie marking prediction method based on the implicit bias and the interest of the friends, marking bias of the users and article preferences of the friends are overall considered, and the recommendation quality is improved.

The invention discloses a method and device for recommending a sticker, a storage medium and a terminal device. The method includes acquiring a sample image set, and screening a figure picture with the sticker from the sample image set; inputting the figure picture with the sticker into a training model for training, so as to obtain a sticker recommendation model; when it is detected that the photographing function of the attached sticker is opened, acquiring a figure image in the preview image, and determining the recommended sticker, and displaying a sticker recommendation list by means of asticker recommendation model. According to the technical scheme, the effect of improving the recommendation quality of the sticker can be achieved.

The invention discloses a collaborative filtering recommendation method of user rating neighborhood information based on a fuzzy mechanism. The method has the technical scheme that the method comprises the following steps of: 1, obtaining rating information of a user on a project, and creating a rating matrix; 2, calculating the user rating membership according to the rating matrix, and calculating the contribution of the project to the similarity according to the project context information; 3, building user like and unlike similarity according to the rating membership and the similarity membership; 4, carrying out similarity value reduction on users with the small rating number, and building user Jnum similarity; 5, building user final similarity according to the user like and unlike similarity and the user Jnum similarity; 6, selecting first K users with the highest similarity values as reference neighborhood users according to the final similarity, and completing the prediction on target users. Experiment simulation results show that the method provided by the invention has higher recommendation quality than a conventional collaborative filtering algorithm, and can be used for recommending interested projects for the users.

The invention discloses an improved collaborative filtering recommendation method based on user characteristics. The method comprises the following steps that according to the mean opinion score of all users, an original user opinion score range is amended, according to the mean value of the modified user opinion score range and the mean opinion score of all users, original user opinion scores are amended, and after normalization processing, user opinion scores are obtained; according to the user opinion scores obtained after normalization processing, adjustment and normalization processing are conducted on the mean value of score difference values, through the combination with an original Jaccard similarity coefficient, the improved Jaccard similarity coefficient is obtained, and the similarity of the user opinion scores is obtained; according to the gender, age and job characteristic information of users, the similarity of user attributes is calculated; the similarity of the user opinion scores and the similarity of user attributes are combined to serve as the final user similarity, and nearest neighbor computing is conducted, and a recommendation list is generated. The method improves the recommendation quality of the traditional user-based collaborative filtering algorithm and reduces the influence of the data sparseness problem to a certain extent.

The invention discloses a recommendation system noise filtering method based on information entropies and fuzzy C-means clustering. The method comprises steps that first, user historical scoring dataof a target recommendation system is collected and arranged; second, Monte Carlo stochastic simulation is utilized to construct sub data sets of the user scoring data, a recommendation algorithm is utilized to acquire information entropies and recommendation precision of different sub data sets; third, the information entropies are classified according to uncertainty levels, recommendation precision is classified according to recommendation precision levels, and an empirical model is constructed to determine the potential natural noise data proportion; fourth, fuzzy clustering analysis on allthe user scoring data sets is carried out, and noise data is identified and deleted; and fifth, a recommendation algorithm operates for all the scoring data sets, and a recommendation precision indexis utilized to evaluate recommendation quality. The method is advantaged in that quantization measurement of the user scoring information can be realized, and the proposed natural noise data filteringtechnology has certain universality and portability.

The invention discloses an LBS-oriented individual recommendation method based on the Markov prediction algorithm. Data filling is carried out by using the Slope One algorithm; therefore the sparsity problem is solved; the Markov prediction method and collaborative filtering recommendation are combined; contexts are classified according to the scores of interest points generated by users; the context transfer conditions of the users are recorded; time observation sequences are formed according to the context transfer conditions of the users; further, the context information of the users at the next moments of the users are predicted by using the Markov prediction method; in actual life, the geographic positions of the contexts have great influence on the transfer of the interest points of the users; in the process of calculating the similarity of the user contexts by using the collaborative filtering recommendation related techniques, the influence of the geographic positions on the similarity is fully combined; therefore, the quality of recommending the contexts to the users is greatly improved; this shows that compared with the traditional recommendation algorithm, the improved recommendation algorithm of the invention has great advantages and helps to improve the recommendation quality.

The embodiment of the invention discloses an information flow recommendation method and a device based on deep reinforcement learning, equipment and a medium, and relates to the technical field of information processing. The method comprises the steps of collecting a historical click sequence of a target user; calling a preset actor neural network and a preset critic neural network; generating a user recommendation list, and displaying the user recommendation list to the target user to obtain feedback result data and a new historical click sequence generated after feedback; calculating a timedifference error; updating parameters in the critic neural network and the actor neural network; and generating a new user recommendation list, and displaying the new user recommendation list until feedback result data of the target user for the new user recommendation list and a new historical click sequence generated after feedback cannot be obtained. According to the method, the interactivity between the recommendation system and the user is enhanced, the feedback of the user is utilized in real time, the recommendation engine can be continuously optimized, the recommendation quality is improved, the user experience is improved, and the user is effectively attracted to remain.

The invention discloses an asymmetrical weighing similarity based collaborative filtering recommendation method and system. The method includes: determining a user similarity asymmetrical weighing factor according to the proportion of user common scoring items; calculating the similarity among users having the common scoring items through the cosine similarity measurement method and the mean square error measurement method according to the user similarity asymmetrical weighing factor; performing fitting calculation on an original similarity matrix of a user through the matrix decomposition gradient descent method to obtain users who do not have the common scoring items; selecting K users having the maximum similarity as neighbors of a target user according to the similarity among the users having the common scoring items or not having the common scoring items, and predicting scores of user for items which have not being scored according to data of the k neighbors; and generating recommendation items of the target user. The method is accurate in calculation of user similarity and is high in recommendation quality, and can be widely applied to the technical field of recommendation.

The invention relates to a user self-similarity-based multi-model combination movie recommendation method and belongs to the technical field of data science and data mining. According to the method, results obtained by collaborative filtering and content filtering are fused dynamically based on the self-similarity of score information of users for movies to obtain a recommendation result. When multi-value attributes of the movies are combined with classification tree models in machine learning, it is proposed that attribute values of a movie are separated to form independent eigenvectors and the multi-value attributes are well combined with the classification tree models. Compared with an existing method, the user self-similarity-based multi-model combination movie recommendation method has the advantages that the recommendation result obtained by fusing the results obtained by collaborative filtering and content filtering dynamically based on the self-similarity of the score information of the users for the movies better meets the user demand, and the recommendation quality is higher.

The invention discloses a position recommendation method and device, a server, a terminal and a storage medium, and belongs to the technical field of Internet. The method comprises the following stepsof in a plurality of geographic areas obtained by dividing a preset geographic range, determining candidate geographic areas according to n acquired meeting departure positions; in the candidate geographic areas, determining a target geographic area with the recommendation level meeting the lowest level requirement; determining at least one reference position in the target geographic area; and determining a recommendation position of the party place according to the at least one reference position. According to the method, the problem that the accuracy of recommending the geographic positionby the server is low can be solved; due to the fact that the recommendation position is determined according to the reference position in the target geographic area with the higher recommendation level, the higher probability of the recommendation level of the recommendation position is, the better the recommendation quality of the geographic position with the higher recommendation level is, and the recommendation quality of determining the recommendation position by the server can be improved.