54 results about "Probabilistic matrix factorization" patented technology

A stream of ordered information, such as, for example, audio, video and / or text data, can be windowed and parameterized. A similarity between the parameterized and windowed stream of ordered information can be determined, and a probabilistic decomposition or probabilistic matrix factorization, such as non-negative matrix factorization, can be applied to the similarity matrix. The component matrices resulting from the decomposition indicate major components or segments of the ordered information. Excerpts can then be extracted from the stream of ordered information based on the component matrices to generate a summary of the stream of ordered information.

The invention relates to a computer aided newsmaker retrieval method based on multimedia analysis. The method comprises the following steps: data preprocessing is performed on a news picture; the multimodality fusion figure relation is initialized; event relation is initialized, a multi-relation probability matrix decomposition model is provided, the potential relations are mined, the correlations between the newsmakers and news events and the query keyword are sorted according to the query keyword submitted by a user and the reconstructed relations; and a result browsing interface is retrieved, namely the name of a person which is submitted to a computer by the user, is used as the retrieval keyword, a relation view using the query person as the center and a related news event list view are provided, and the retrieval result is fed back to the user.

The invention discloses a collaborative filtering method on the basis of a scene implicit relation among articles. The collaborative filtering method comprises the following steps of: 1, extracting scores of the articles in different scenes from original score data and establishing an article-scene score matrix; 2, decomposing the article-scene score matrix by a matrix decomposition method to obtain an implicit factor matrix of the articles; 3, establishing a scene feature vector for each article by using the obtained implicit factor matrix of the articles so as to calculate the similarity among the articles by utilizing a Pearson correlation coefficient and establish an article implicit relation matrix; and 4, integrating obtained article implicit relation information into a probability matrix decomposition matrix to generate a personalized recommendation. According to the invention, scene information can be sufficiently utilized to mine the implicit relation information among the articles, and the recommendation is generated by utilizing the implicit relation among the articles; the collaborative filtering method has high expandability for the scene information, and a candidate scene set can be regulated according to the application requirements; and the accuracy and the personalization degree of the recommendation can be effectively improved.

The invention is suitable for the field of social network information recommendation and provides a probability matrix decomposition-based community trust recommendation method and system. The method comprises the steps of obtaining behavior data of users, and obtaining a community set; building a community data model according to user information, a user relationship, project information required to be recommended and the community set, and obtaining a trust relationship among the users; building a probability matrix decomposition model according to the trust relationship, solving the model, and performing calculation to obtain recommendation information; and recommending required project information in the recommendation information to corresponding users according to a predetermined recommendation rule. According to embodiments of the method and the system, a community relationship of the users is introduced in the trust relationship, and the users in the same community are distinguished from the users in different communities, so that the trust relationship can reflect a real relationship among the users more accurately, the situation that a one-sided interactive relationship is regarded as the trust relationship is avoided, and the accuracy of a trust-based recommendation algorithm is greatly improved.

The invention discloses a group-oriented project recommendation method based on joint probability matrix decomposition. The group-oriented project recommendation method is characterized by comprising the following steps: (1), representing project scoring information of a user and information that the user belongs to a group by using a triple; (2), calculating the user relevance based on the triple of the information that the user belongs to the group; (3), implementing a joint probability matrix decomposition method based on the user relevance so as to obtain a user characteristic matrix and a project characteristic matrix; (4), calculating by utilizing an average strategy so as to obtain a group characteristic matrix; and (5), obtaining previous N projects having the highest project predication scoring in each group so as to obtain a recommendation list. According to the invention, the user relevance based on group information is integrated in probability matrix decomposition; joint probability matrix decomposition based on the user relevance is implemented; furthermore, the group characteristic matrix is obtained by calculation through the average strategy; therefore, a group-oriented recommendation result is obtained; and more accurate individual services can be provided for the group to a certain degree.

The invention discloses a personalized recommendation method based on probability matrix decomposition combined with similarity. The method comprises the specific steps that S1, an article information and historical score database is established; S2, a similarity matrix of users and articles is generated; S3, row vectors of the matrix are arranged according to a descending order; S4, an objective function is generated based on a probability matrix decomposition model; S5, a final potential feature matrix of the users and the articles is generated; S6, a prediction score matrix is generated according to the final potential feature matrix of the users and the articles; and S7, personalized recommendation is made to the users. According to the method, prediction scores can be more approximate to true scores of the users according to the condition that potential feature vectors of the users are relevant to user potential feature vectors with high similarity, and therefore the recommendation accuracy of a current recommendation system is improved.

The invention relates to a mining method of computer data for recommendation systems and belongs to the technical field of computer data processing. The method includes that user preference matrixes and service project preference matrixes are initialized in a main server of a computer, row vectors of preference matrixes input by a user are distributed to a plurality of mappers in the computer, all mappers calculate sub-directions of gradient directions of the user preference matrixes and the service project preference matrixes respectively, calculated results are sent to a simplifier in the computer, the simplifier receives and accumulating the sub-directions of the gradient directions, and the user preference matrixes and the service project preference matrixes are updated according to gradient direction matrixes of the user preference matrixes and the service project preference matrixes. According to the method, the existing probabilistic matrix factorization (PMF) algorithm is modified, large-scale data processing capabilities are improved; and preference matrixes are stored by data storage structures of key-value pairs, so that occupied storage space is small, and reading speed of data are rapid.

The invention provides a probability matrix decomposition recommendation method. The method provided by the invention is based on explicit and implicit feedback information and combined with a probability matrix decomposition technology at the same time so as to help a user to score the commodity. The method comprises the steps of firstly carrying out probability matrix decomposition on a user trust relationship matrix, then carrying out probability matrix decomposition on a commodity score matrix of the user and the implicit feedback information, solving an implicit user characteristic matrix, an implicit commodity characteristic matrix and an implicit feedback information characteristic vector by integrating results of the two times of decomposition, and finally calculating a recommendation score of the user. The method provided by the invention can help the user to recommend a score of the commodity by well utilizing an online social network, relieves a data sparse problem and a cold start problem in the social network, and has an excellent recommendation effect. Meanwhile, the probability matrix decomposition recommendation method can be applied to a recommendation system with a large-scale data set.

The invention discloses a weighted trust relationship-based probability matrix decomposing and recommending method. The weighted trust relationship-based probability matrix decomposing and recommending method comprises the following steps: acquiring a user item score information matrix and an inter-user trust relationship score matrix and filling a trust information matrix with trust data; for theuser item score information matrix, calculating the similarity of a target user and each user according to a similarity calculation formula; fusing the acquired inter-user similarity into trust scoredata to acquire weighted trust score data; combining the acquired weighted trust score data with a probability matrix decomposing method to acquire a probability matrix decomposing and recommending model based on the weighted trust relationship; successfully predicting an item in which the target user is interested according to the acquired recommending model. The weighted trust relationship-based probability matrix decomposing and recommending method mainly aims at a social network site with score information and trust data and is mainly applied to an electronic commerce system; high-qualityand high-accuracy recommendation for the target user is formed effectively.

The invention aims at solving the problem in the prior art, and provides a collaborative filtering method of simultaneously integrating friend characteristics and similar user characteristics. The accuracy of the collaborative filtering method is improved. The method comprises the steps of obtaining a user-object scoring matrix, calculating the similarity between users through a Pearson correlation coefficient and building a user similarity matrix; obtaining a user-user relationship matrix and carrying out normalization processing on an original user-user relationship matrix to obtain a normalized user-user relationship matrix; and integrating the similar user characteristics and the friend user characteristics into the user characteristics of a probability matrix decomposition model according to the user similarity matrix and the normalized user-user relationship matrix, obtaining prediction scores of users on objects according to the probability matrix decomposition model and recommending the objects for the users according to the prediction scores. The collaborative filtering method has the beneficial technical effects that the recommending result is affected by the friend user characteristics and the similar user characteristics, and the accuracy of the collaborative filtering method is improved.

The invention discloses a high-spectral image sharpening method based on probability matrix decomposition, and belongs to the field of remote sensing image processing. The method is characterized in that the method comprises the steps: carrying out the preprocessing of two inputted images based on the hypothesis that a pixel spectrum vector of a high-resolution high-spectral image is just formed by the linear superposition of a few of vectors with the hidden spectrum features according to one low-resolution high-spectral image and one high-resolution high-spectral image and the frequency response matrix, decomposition matrix dimensions and algorithm iteration number, corresponding to the high-spectral images, of a multispectral camera, wherein the low-resolution high-spectral image and the high-resolution high-spectral image are taken at the same height in the same target region at the same time; listing mathematical equations of the two processed images and a to-be-solved high-resolution high-spectral image, and building a Bayesian model; calculating the posteriori probability distribution of the decomposition matrix, and obtaining a matrix with the hidden spectrum features in the decomposition matrix, and solving the mean value of linear superposition vectors corresponding to the two images after preprocessing, thereby obtaining the to-be-solved high-resolution high-spectral image. The method greatly reduces the time consumption of calculation while improving the sharpening precision, and is easy to adjust.

The invention discloses a social recommendation method based on trust relationship implicit similarity. Firstly, trust and trusted vectors of users are obtained through the probability matrix decomposition, wherein the vectors imply direct and indirect association between the users, and accordingly sparse scoring and information in trust data are mined more sufficiently; secondly, trust relationship implicit similarity of the users can be obtained through the probability estimation method, and accordingly estimation precision of trust relationship strength is improved; finally, scoring similarity between the users is comprehensively taken into consideration, and robustness of the algorithm under the data sparsity condition is further improved.

The invention provides a missing data recovery method and device in order to achieve recovery of missing data. The missing data recovery method comprises the steps that multiple sets of data are obtained; a numerical value matrix composed of the multiple sets of data is subjected to probability matrix decomposition; the position of the data missed in the multiple sets of data is determined; the product of the elements, corresponding to the position of the data missed in the multiple sets of data, in a probability matrix decomposition result is obtained to be used as missing data; the obtainedmissing data is recovered to the position of the data missed in the multiple sets of data.

The invention discloses a method for establishing a probability matrix decomposition model based on a node user. The method comprises the steps that firstly, an influencing degree of the user is defined according to influences of a social network on recommendation; secondly, formalized measurement is conducted to the influencing degree of the node user; thirdly, the probability matrix decomposition model based on the node user is established; and fourthly, the probability matrix decomposition model based on the nose user is trained. In comparison with the prior art, the method disclosed by the invention has the advantages that effects of social relations on the recommendation can be enhanced; and prediction accuracy of a recommendation algorithm can be further increased.

The invention relates to a collaborative filtering method for user interest harmony similarity. The method comprises the following steps: decomposing an original scoring matrix; introducing a time-based user interest degree weight function into a Pearson similarity function; and combining with the user evaluation grade trust factor and the user evaluation deviation trust factor to obtain user interest harmony similarity, obtaining a recommendation result by utilizing the original scoring matrix and the user interest harmony similarity, and recommending items meeting requirements to the user. According to the method provided by the invention, the recommendation precision can be remarkably improved compared with a traditional probability matrix decomposition method under the condition that the user scoring data is sparse, and the interested articles can be recommended according to the user in different time periods.

The invention discloses a matrix decomposition recommendation method based on joint clustering. The matrix decomposition recommendation method includes the steps that a user-item rating matrix is constructed; the user-item rating matrix is divided into multiple categories through joint clustering; for the categories obtained after clustering, a probability matrix decomposition method is used for parallely predicting unknown scores of all the categories, and recommendation is carried out according to the predicting scores. Close correlation in the clustering and high precision of the probability matrix decomposition algorithm can be used fully, and for the big data processing problem in the information overload times, recommendation can be conducted at high speed while high precision is ensured.

The invention provides a data compression method and device, aiming at realizing the compression of multiple sets of data including missing data. The data compression method includes the following steps: acquiring multiple sets of data; performing probability matrix decomposition on a numerical matrix composed of the multiple sets of data; and compressing the multiple sets of data by using a probability matrix decomposition result.

The invention discloses a probabilistic matrix factorization(PMF) model realizing fusion of similarity and common rating item quantity and provides a novel probabilistic matrix factorization model capable of fusing similarity and common rating item quantity. According to the model based on the PMF, the inner product of the factor matrix is restrained by the similarity, an appropriate penalty function is built to the common rating item quantity, the penalty function and the inner product of the factor matrix are associated to realize co-constraint to the factor matrix, and thus, quality of the factor matrix is improved. The advantage of the similarity in terms of exploring local relation and the advantage of the PME model in terms of exploring global relation are combined successfully; besides, according to the model, the influence from the common rating item quantity on trustworthiness of the similarity is also taken into consideration. Compared with the previous research, the model has the advantage of higher precision.

The invention discloses a matrix decomposition interest point recommendation method based on geographic position fusion social influence and category popularity. The method comprises the steps of obtaining historical sign-in information of all users and candidate interest points in the position social network, calculating a geographical correlation coefficient of a target user to obtain social network information of all the users, and calculating a social correlation coefficient S (xu, l) between the target user and the interest points; calculating a category popularity correlation coefficientC (yu, l) between the target user and the interest point based on the preference of the candidate interest point category and the popularity of the candidate interest point; fusing the geographic information model, the social correlation coefficient, the category popularity correlation coefficient and the potential features through a probability matrix decomposition method to form an interest point recommendation model, and calculating a recommendation score Recu, l of the target user; and generating recommendation information based on the recommendation score Recu, l, thereby realizing the purpose of personalized recommendation.

The invention discloses a personalized recommendation method based on probability matrix decomposition. The personalized recommendation method comprises the steps of 1, calculating user attribute similarity; 2, calculating the attention similarity of the user to the label and the score similarity of the user to the label, and integrating the two similarities to generate the preference similarity of the user; 3, proportionally adding the user attribute similarity and the user preference similarity into 1 to generate a final user comprehensive similarity matrix; 4, initializing a user implicit feature matrix and a project implicit feature matrix into normal distribution, and fusing a user comprehensive similarity matrix into the probability matrix decomposition model; and 5, iteratively calculating a user feature matrix and a project feature matrix, generating a prediction scoring matrix of the user for the projects, and recommending the first K projects with the highest scores to the user as final recommendation results by adopting a Top-K sorting method. The method has the advantages that the user cold start problem in the recommendation system is relieved, the accuracy of the recommendation algorithm is improved, and the expandability of the algorithm is enhanced.

The invention relates to a matrix decomposition method and device based on convolution attention and an electronic device. The matrix decomposition method based on convolution attention comprises thefollowing steps: expressing a user description document of an article as a word vector matrix; inputting the word vector matrix into a convolutional attention neural network to obtain a hidden factorof the article; obtaining hidden factors of the user by using probability matrix decomposition according to the scoring information of the user to the articles and the hidden factors of the articles;calculating an inner product of a vector for the hidden factors of the article and the hidden factors of the user, and establishing a prediction scoring matrix according to the inner product; and optimizing parameters of the convolutional attention neural network and the probability matrix decomposition by using an error back propagation algorithm. The matrix decomposition method based on convolution attention is good in interpretability, can effectively reduce the problem of cold start of articles, can improve the problem of data sparsity, and improves the accuracy of score prediction.

The invention relates to a hybrid recommendation algorithm based on a deep neural network and probability matrix decomposition. The method comprises steps of S10, collecting and preprocessing information; S20, establishing a deep neural network model and a probability matrix decomposition model; S30, obtaining a potential feature vector fusing real information of the user and the project accordingto the three steps; and S40, performing personalized recommendation on the user by utilizing the feature vector. The invention discloses a hybrid recommendation algorithm based on a deep neural network and probability matrix decomposition. The beneficial effects of the invention are as follows: the method is based on previous research; the user scoring matrix data is fully utilized; and meanwhile, feature extraction is performed on the user description information and the project description information by using a deep neural network to generate a user and project real feature set includinguser preference features, and performing iterative optimization processing on the probability matrix model and the features generated by the deep neural network by using maximum posteriori estimation.

The invention provides a collaborative filtering recommendation algorithm and device based on similarity and similarity credibility. The algorithm restricts the included angle cosine of a potential eigenvector through a similarity relation on the basis of a probability matrix decomposition model, and influences the constraint weight through the similarity credibility. Due to the constraint of similarity credibility, the algorithm effectively weakens the influence of sparsity on similarity relations, enhances the learning ability of important similarity relations, and alleviates the sparsity problem of scoring. Experimental results show that the proposed algorithm is superior to other similar algorithms. The next work is to apply this algorithm to the actual recommendation system to overcome the data sparsity problem and the poor recommendation effect of a conventional recommendation system in the practical application.

The invention discloses a probability matrix decomposition cold start recommendation method fusing attributes and semantics. The method comprises the steps of firstly, extracting user attribute information, project attribute information, project text information and user scoring information from a database; and modeling the attribute information and the semantic information by utilizing linear regression to predict potential characteristics, and taking a predicted value as a prior probability of probability decomposition, thereby realizing fusion of the attribute information and the semantic information into probability decomposition of a scoring matrix. According to the method, attribute information and semantic information can be effectively fused into probability matrix decomposition, the ubiquitous problems of cold start and sparsity in a recommendation system are solved, and the method has higher accuracy and low algorithm complexity and is suitable for processing large-scale data.

A time-sensitive personalized recommendation method based on probability matrix decomposition belongs to the field of context-aware personalized recommendation, comprising: constructing a user-contextscoring matrix by using the user's rating information and context information of the movie and the movie category: constructing a user-context scoring matrix according to the original user-movie scoring matrix and additional context information and movie category information, performing matrix decomposition then to obtain a user feature vector containing the context information, then solving thecosine similarity between the two users, selecting a certain number of users with higher similarity as the neighbor users, and integrating the influence of the neighbor users into the probability matrix decomposition for scoring prediction, thereby making personalized recommendations based on the level of the rating forecast. The invention is applicable to the fields of movie recommendation, e-commerce website products, digital library book recommendation, Internet advertisement placement and the like which require personalized recommendation.

The invention provides a graph model-based QoS prediction method applied to cloud service recommendation. The method comprises the steps that a full-graph model containing multi-source information isconstructed, and the full-graph model comprises nodes representing users and services and edges using the similarity between the users, the similarity between the services and the similarity between the users and the services as weights; the full-graph model is segmented into a plurality of sub-graph models; a probability matrix decomposition algorithm for respectively optimizing the full-graph model and the sub-graph model is used for obtaining QoS of global and local prediction; and adaptive fusion processing is carried out on the QoS of global and local prediction to obtain the QoS of finalprediction. According to the method provided by the invention, the influence of multi-source information on QoS is fully considered, and local and global features are adaptively fused, so that the QoS prediction precision is improved. According to the method, the missing QoS value can be accurately predicted, then the sparse matrix is filled, the density of the matrix is improved, and the problemof QoS sparsity in the field of cloud service recommendation is solved to a certain extent.

The present invention is applicable to the field of computer technology, and provides a recommendation method, device, equipment, and storage medium. The method includes: acquiring the user's historical rating data, items to be rated, and the text content of the item to be rated, and according to the user's historical rating data, The items to be rated and the text content of the items to be rated are trained on the preset stacked noise reduction autoencoder and the preset probability matrix decomposition model to obtain the user feature matrix, item feature matrix, and corresponding item hidden features and user hidden features. Features, according to the user feature matrix, item feature matrix and the corresponding item hidden features and user hidden features, calculate the user's predicted score for the item to be rated, generate a recommendation list according to the predicted score, and output the recommendation list to the user, so as to recommend items to the user Combining item features and user features at the same time can effectively improve the accuracy of recommendation, and then improve the efficiency of item recommendation.

The invention discloses a probability matrix decomposition recommendation method based on user context coupling similarity. The method comprises the steps of firstly, constructing a user relationshipmatrix in combination with the coupling similarity of user context information and user credibility; combining the user trust relationship matrix and the user project matrix to carry out probability decomposition to learn a feature vector; and finally finishing recommendation through trusted users of new users. The invention aims to provide the probability matrix decomposition recommendation method based on user context coupling similarity. The user context information coupling similarity serves as a special trust relationship, a user trust relationship matrix is constructed through the coupling similarity between the user context information and the user trust degree, and then potential feature vectors are learned through probability decomposition in combination with a user scoring matrix, so that recommendation is achieved. The problems of cold start and sparsity which widely occur in a recommendation system are solved.

The invention provides a probability matrix decomposition recommendation method based on user comments, and particularly relates to the field of data mining. According to the method, probability matrix decomposition is enhanced through the user comment text, and the problems of data sparsity and cold start in a recommendation algorithm are solved. The method comprises the following steps: firstly, selecting user comments capable of representing project semantics as auxiliary information, and carrying out Word2vec word vector training and preprocessing on a comment text by utilizing an open source tool to obtain a project tag; and then a stack type noise reduction auto-encoder in deep learning is used to extract item features in the tag, and the stack type noise reduction auto-encoder converts sparse high-dimensional data into low-dimensional data, so that the robustness of a recommendation algorithm is enhanced. and finally, probability matrix decomposition is performed on the user score matrix and the extracted item matrix, and the item feature matrix extracted from the user comment text enhances the probability matrix decomposition, so that the recommendation effect is greatly improved.