78results about How to "Improve recommendation effect" patented technology

The invention provides a recommendation strategy generation method based on big language model enhancement and related equipment. The method comprises the steps of obtaining a recommendation task; inputting the recommendation task into an intelligent agent, and obtaining a recommendation strategy output by the intelligent agent; wherein the intelligent agent is obtained by training based on a sample recommendation task and a large language model, and the large language model is used for predicting user behaviors based on the sample recommendation task. According to the method, the big language model is combined with the reinforcement learning agent, so that the problem that an accurate recommendation strategy is difficult to quickly provide by a traditional method is effectively solved, and the efficiency and the adaptability of a recommendation system are improved while the accuracy and the individuation degree of the recommendation strategy are improved.

The invention discloses a privacy protection-oriented personalized federal recommendation system and method, and relates to the field of federal recommendation systems. According to the method, an article preference embedding vector is constructed for each client, and a global article embedding orthogonal mapping matrix and a double-layer perceptron used for generating an article embedding mixed factor matrix are trained, so that the global article embedding matrix and the article preference embedding vector of a user are fused; a mixed article embedding matrix which is more personalized and finer in granularity is constructed for the client, and the mixed article embedding matrix is input into a personalized recommendation prediction perceptron constructed for the client to obtain a recommendation result, so that the final recommendation effect is improved; when the client side uploads data to the central server, local differential privacy encryption is carried out; according to the technical scheme, negative sampling is carried out on each client side, the positive samples and the negative samples are mixed and then jointly used for training, the training effect is improved, meanwhile, the server cannot distinguish the historical interaction situation between the client sides and the articles according to the update data uploaded by the client sides, and user privacy is protected.

The invention discloses a multi-modal robust recommendation method and system based on self-supervised learning, and the method comprises the steps: extracting the modal features of each modal of a user and an article from original data as initial embedded vectors, and constructing a user-article heterogeneous graph and an article-article isomerism graph under each modal; performing multi-modal feature coding on the initial embedded vector of each modal, and performing multi-modal feature fusion on the enhanced final embedded vector of each modal; and performing multi-modal feature alignment to extract a final user embedded vector # imgabs0 # and a final article embedded vector # imgabs1 #, calculating to obtain a final prediction score of the user on the articles, sorting the final prediction score, selecting specified first K articles as a recommendation list recommended to the user, and outputting the recommendation list. The invention aims to eliminate the influence of inherent noise in the multi-modal data and improve the recommendation effect of the whole multi-modal recommendation system.

The invention discloses a nursing equipment recommendation method and system based on user historical behaviors, and relates to the technical field of equipment recommendation. Behavior patterns of users are identified, the users are divided into different groups based on a clustering result, user behavior data similarity is described by using a semi-variance function to form a semi-variance model; after a corresponding vacancy filling model is selected based on semi-variance model features, the selected vacancy filling model is trained according to the clustered user behavior data so as to estimate behavior data missing in time or space, the weight of each feature data is distributed in combination with the feature data of the user behavior data and a prediction result of the vacancy filling model, and the weight of each feature data is calculated. And setting a recommendation rule by using a multi-factor weighting method, and generating a nursing equipment recommendation list based on the recommendation rule. The recommendation system can utilize known data points to estimate missing values, so that the problem of data discontinuity or missing is solved, the space-time gap of user behavior data is filled after user behaviors are predicted, and the recommendation effect and accuracy are effectively improved.

The invention relates to the technical field of user personalized content pushing, in particular to a user personalized content pushing system based on big data and a promotion method. The method comprises the following steps: collecting user behavior and preference data through multiple channels; dynamically constructing user portraits, and enriching the user portraits in combination with multi-dimensional data; based on user portraits and content features, a recommendation model is constructed by using a deep neural network, and user preferences are accurately predicted through multiple hidden layers; through a push model and a multi-objective optimization method, the probability of pushed content is calculated in combination with indexes such as a click rate and a conversion rate, the push frequency and content selection are optimized, and the correlation of the pushed content and the user satisfaction are ensured; by monitoring user feedback in real time and adjusting a pushing strategy, intelligent decision and personalized recommendation of content pushing are realized, and long-term accurate pushing and improvement of user experience are ensured. According to the method, the accurate user portrait is constructed through multi-data source collection and processing, so that personalized content pushing is realized.

The invention discloses a text representation alignment sequence recommendation method based on decoupling contrast learning, and belongs to the technical field of sequence recommendation and deep learning. The method comprises the following specific steps: constructing a text-based description data set for sequence recommendation; creating text semantic encoders of the articles and the sequences, and respectively generating corresponding text semantic vector representations; through unsupervised and supervised comparison of learning tasks, learning decoupling representation of sequences and target objects, and realizing bidirectional semantic alignment between the learning tasks and the target objects; according to the method, the overall semantics of the sequence and the features of the single object are effectively separated by decoupling the representation of the sequence and the target object, and the high-quality semantic representation of the interaction sequence and the target object is realized, so that the accuracy and effect of sequence recommendation are remarkably improved. According to the method, decoupling contrast learning is innovatively applied to sequence recommendation, the method can be widely applied to scenes such as e-commerce recommendation and content recommendation, and the method has important practical application value.

The invention relates to an intelligent product recommendation method and system based on a large model. The method comprises the steps of obtaining user information of a user to be recommended and product information of a product to be recommended; extracting user semantic features, user behavior features and natural language understanding features of the user information and product description features and product category features of the product information by using the large model, and performing feature fusion to obtain fused features; and obtaining a target recommendation score of each product relative to the to-be-recommended user, and recommending the product corresponding to the target recommendation score higher than a preset score to the to-be-recommended user as a target product. According to the method, the Transform model, the GNN model and the LLM model are combined, analysis and feature fusion of the user information and the product information are achieved, the extraction precision of the user semantic features and the product description features is improved, the user behavior features and the product category features can be effectively captured, and therefore the recommendation accuracy and the personalized level are improved.

The invention discloses a similarity and popularity combined collaborative filtering method. The method comprises popularity calculation and popularity and similarity combined collaborative recommendation. The invention provides an article popularity iterative calculation method introducing high-order information, and constructs a contrast loss function to realize effective fusion of similarity and popularity. Experimental studies performed on three common data sets show that compared with a mainstream collaborative recommendation method, the model provided by the invention has obvious advantages in the aspects of recommendation performance and operation efficiency.

The invention relates to the field of artificial intelligence, and provides an information recommendation method and device, equipment and a storage medium. The method comprises the steps that a recommendation model conducts recommendation estimation on multiple pieces of candidate recommendation information obtained through retrieval, target recommendation information recommended to a target object is determined based on corresponding information recommendation estimation values, and the recommendation model is obtained by adopting a loop iteration mode and based on first training samples of classification targets with the positive and negative sample number greatly different. In each round of iterative training, each first training sample is divided into a common sample or a difficult sample, a first weight value with a relatively low weight value is given to a target weight of the common sample, a second weight value with a relatively high weight value is given to a target weight of the difficult sample, and the second weight value is in positive correlation with the learning difficulty of the corresponding difficult sample; based on the first recommendation estimated value of each training sample and the target weight, model parameters are adjusted, and a model capable of correctly distinguishing positive and negative sample interfaces is adopted, so that the recommendation accuracy is improved.

The invention relates to a shopping center shop recommendation method and device, equipment and a medium, and the method comprises the steps: predicting interested shops of a target old user in a current time period through employing a long and short-term memory network model trained by user behavior time sequence data, adopting a preset collaborative filtering algorithm to calculate and determine a second similarity between the target old user and other users according to the interested shops of the target old user in the current time period, so as to determine neighbor users of the target old user; predicting an estimated score value of a target old user for an unscored shop according to the score value of the neighbor user for each shop, and calculating and determining a second recommendation degree corresponding to each shop according to the estimated score value, the actual distance between the target old user and the shop and the comprehensive recommendation value, and displaying each shop to a user interface of the target old user according to the second recommendation degree from high to low so as to complete shop recommendation of the shopping center. According to the invention, the traffic and user stickiness of a shopping center can be improved.

The application provides a dynamic increment-based graph information gate aggregation knowledge graph recommendation method, solves the data instability caused by random sampling and the noise fusion problem caused by double diffusion in the knowledge graph convolution recommendation method. The method comprises: dynamically selecting Top-K adjacent edges and adjacent point fusion for the current node in the knowledge graph (K is a positive integer of 10-20) in sequence. On the one hand, the matching score value of the current node, user embedding, adjacent edge and adjacent point is calculated by using the attention mechanism, the high correlation of the adjacent edge, adjacent point graph information and the candidate node of the next step is selected by using the score value, on the other hand, the selected graph information is adaptively adjusted and fused by using the gate mechanism, the adjacent information is accurately integrated into the current node embedding, and then the candidate node is diffused in sequence. Finally, through the operation of the adjacent fusion, the current node information and the adjacent graph information can be directly fused once, so that the noise information in the double diffusion and multiple fusion graphs is prevented.

The invention relates to the technical field of private domain traffic, in particular to a private domain traffic guiding method and system based on collaborative filtering, and the method comprises the following steps: S1, collecting user behavior data: collecting the behavior data of a user in a private domain platform in real time through client equipment; s2, constructing a user portrait: extracting the interest preference of the user, generating a user interest label, and forming the user portrait; s3, interest extension and cross-domain recommendation: constructing a relation graph between the user and the content, identifying association between the user interest node and the cross-domain content, calculating the edge weight between the user and the content node, and extending the interest boundary of the user in the private domain traffic; and S4, recommended content pushing: pushing cross-domain recommended content to the user, and meanwhile, optimizing the display sequence of the content in combination with the situation information of the user. According to the invention, the diversity and personalization of the recommended content can be enhanced, and the user experience and the diversity of the recommendation effect are further improved.

The invention relates to an object recommendation interaction method and device, electronic equipment and a storage medium. The method comprises the steps that recommended media content corresponding to a target object is displayed on a preset page; in the recommended media content display process, displaying a first part of picture in a first preset recommended picture corresponding to the target object in a preset area in a preset page, and displaying a second part of picture in the first preset recommended picture in the preset page; in response to a preset interaction operation triggered by the target account, moving the second part of the picture to a first preset sub-region in the preset region to enable the preset region to display a complete first preset recommended picture; and skipping to an associated page corresponding to the target object under the condition that the complete first preset recommendation picture is displayed. By utilizing the embodiment of the invention, the presentation style of the object recommendation picture can be enriched, the effectiveness and interestingness of user interaction are improved, the user interaction experience is improved, and the object recommendation effect and the object recommendation efficiency are improved.

The invention provides a personalized recommendation method for multi-behavior association modeling and interest boundary perception. Comprising a transformer-based multi-behavior sequence modeling module, a time-aware graph aggregation module and a user interest boundary prediction module, and a complete system is realized through training. An interaction sequence generated by a user before the current moment and a corresponding behavior type are embedded and then pass through the transformer-based multi-behavior sequence modeling module to obtain a behavior and sequence sensing node representation, and then through the time sensing graph aggregation module, a representation of a one-hop center node is obtained through aggregation according to the distance of a time interval. Stacking L layers to expand the user or project interaction diagram to L hops to obtain user node and project node representations; and finally, performing loss function optimization through the user interest boundary prediction module after the user node and the project node representation are subjected to MLP to obtain a prediction result of the user interest, and realizing personalized recommendation.

The application discloses a kind of individual and group-oriented behavior internet modeling and prediction recommendation method, the method includes the following steps: step 1: establish the behavior preference model of behavior internet based on hypergraph;Step 2: extract the relationship between behaviors based on the model of step 1;Step 3: discover the behavior mode of user based on the behavior relationship of step 2;Step 4: predict the behavior trend of user based on the behavior relationship of step 2 and the behavior mode of step 3;Step 5: personalized service recommendation is carried out for user based on step 2, step 3, step 4.The model of behavior internet is proposed in the application, and the personalized behavior internet is obtained by cause-effect inference and relationship discovery, and the recommendation is carried out by using the deep learning method based on external knowledge, which improves the recommendation efficiency and effect for individual and group, overcomes the shortcomings that the traditional recommendation method based on deep learning is difficult to fully utilize personalized knowledge, and realizes the personalized and efficient recommendation of service.

The invention provides a text generation method and related equipment. The method comprises the following steps: performing text recognition on a sample video about a sample object to obtain sample information of the sample object and a corresponding initial text; performing abnormal text correction and text screening on the initial text to obtain a screened text; performing model training based on the sample information and the screened text to obtain a text generation model; the text generation model generates a target text used for recommending the target object based on the target information of the target object.

This application provides a video recommendation method, apparatus, device, and storage medium, relating to the field of communication technology. The method includes: upon receiving a video acquisition request, acquiring a candidate video list; acquiring the interaction probability between the user and each candidate video in the candidate video list, the interaction probability being determined based on the user's reconstructed embedding vector and the candidate video's reconstructed embedding vector; the user's reconstructed embedding vector being determined based on the user's embedding vector, the user's feature vector, and a pre-trained vector reconstruction model; the candidate video's reconstructed embedding vector being determined based on the candidate video's embedding vector, the candidate video's feature vector, and the vector reconstruction model; simulating the input of sample data during cold start when training the vector reconstruction model; and determining the target video to recommend to the user from the candidate video list based on the interaction probability between the user and each candidate video in the candidate video list. This method can accurately recommend videos to users during cold start, improving the video recommendation effect during cold start.

The invention is suitable for the field of artificial intelligence, and provides a transportation tool matching recommendation method and system based on a knowledge graph, and the method comprises the steps: converting a triple into a text form through employing an improved knowledge graph embedding model KGE, carrying out the coding of an input text through a BERT model, and obtaining the context representation of each Token; capturing global semantic information of the entities and the relationships based on the aggregation representation of the triple to obtain embedded representations of the entities and the relationships; integrating and extracting features of the heterogeneous information, and performing dimension reduction processing on feature data of the heterogeneous information by using principal component analysis (PCA) to obtain embedded representation of the heterogeneous information; the method comprises the following steps: performing feature dimension reduction on entities, relationships and heterogeneous information based on an embedded layer of a recommendation system algorithm NFM model, capturing a second-order interaction relationship between dimension reduction features through bilinear interaction operation, learning high-order feature interaction by using a feed-forward neural network FNN, outputting a matching probability of goods and a transportation tool, and obtaining a matching result of the goods and the transportation tool; and based on the matching probability, performing matching recommendation of the goods and the transportation tool.

This application provides a video recommendation method, apparatus, device, and storage medium, relating to the field of communication technology. The method includes: upon receiving a video acquisition request, acquiring a candidate video list; acquiring the interaction probability between the user and each candidate video in the candidate video list, the interaction probability being determined based on the user's reconstructed embedding vector and the candidate video's reconstructed embedding vector; the user's reconstructed embedding vector being determined based on the user's embedding vector, the user's feature vector, and a pre-trained vector reconstruction model; the candidate video's reconstructed embedding vector being determined based on the candidate video's embedding vector, the candidate video's feature vector, and the vector reconstruction model; simulating the input of sample data during cold start when training the vector reconstruction model; and determining the target video to recommend to the user from the candidate video list based on the interaction probability between the user and each candidate video in the candidate video list. This method can accurately recommend videos to users during cold start, improving the video recommendation effect during cold start.

The application discloses a user sensitive information protection method based on a data confusion technology, and comprises the following steps: step S1, generating a sensitive attribute association table based on the association between recommended items and user gender characteristics; step S2, adding confusion ratings to an existing user-item matrix through a sampling strategy according to the sensitive attribute association table obtained in step S1, and constructing a user-item matrix after adding the confusion ratings; step S3, recording the number of confusion ratings added in step S2, applying a removal strategy, and finally generating a confusion matrix after applying the removal strategy; the removal strategy is that the same number of confusion ratings are deleted to maintain the original data size; a user is randomly selected, and when the rating number of the user reaches a set threshold, the items with strong association of the opposite sex are deleted from the rating history of the user. The application can mislead attackers to conduct gender attribute inference, realize user privacy protection, and does not affect the existing personalized recommendation quality of the user.

The invention discloses a multi-model denoising method and system for a reinforcement learning recommendation system, and the method comprises the steps: obtaining noisy interaction data, and carrying out the prediction of the noisy interaction data through employing a generalized matrix decomposition model and other two matrix decomposition models as prediction models; predicting model parameters are optimized by minimizing KL divergence of fitting real user preference distribution among a plurality of predicting models and maximizing likelihood of observation data under given real user preference conditions; the optimized prediction model serves as a teacher model, and the teacher model is trained through the noise reduction positive sample noise and the noise reduction negative sample noise to obtain a de-noised teacher model; constructing a student model, and defining a target loss function; according to the method, the denoising algorithm is used for training the denoising user model, the denoising user model is used for interacting with an online reinforcement learning strategy, feedback better conforming to the real preference of the user is provided, and the noise robustness and the recommendation performance of the reinforcement learning recommendation method are improved.

The invention discloses an electric power material commodity recommendation method and device based on an improved deep interest network. According to the method, the problem of inaccurate commodity recommendation caused by lack of sufficient interaction data of new articles or a long tail effect in an electric power material recommendation system is solved. By introducing an instance attention mechanism and a channel attention mechanism, feature expression is enhanced, so that a recommendation system can make accurate commodity recommendation under the condition of a small number of samples, and the method is particularly suitable for commodity recommendation scenes with professional characteristics such as electric power materials; and the smoothness of the training process is dynamically adjusted by using the Cauchy loss, so that the ability of the model to capture the deep characteristics of the high-dimensional incomplete data is enhanced, and the model is further endowed with stronger representation ability and robustness.

The invention relates to a strategy model training method and device, a content sorting method and device and computer equipment. The method comprises the following steps: determining a plurality of sample user state representations based on user information obtained in a historical time period; determining a value label corresponding to each sample user state representation based on a plurality of recommended contents in a historical time period; respectively inputting the plurality of sample user state representations into a trained model, and determining sample responses of the sample user state representations for the plurality of recommended contents in the historical time period according to an output result of the trained model; and optimizing model parameters of the trained model according to the plurality of sample user state representations, the plurality of sample responses and the plurality of value labels to obtain a strategy model. By adopting the method, the recommendation efficiency and the recommendation effect can be improved.

The present invention relates to the field of intelligent workbench search optimization technology, and in particular to an intelligent workbench search optimization method and system based on multi-scenario business. The method comprises the following steps: collecting user behavior data of platform users to obtain user behavior data; performing behavior time series analysis on user behavior data to obtain user behavior sequence; constructing a user tag library according to the user behavior sequence to obtain a user tag library; extracting content information from the workbench and constructing a content feature vector library to obtain a content feature vector library; constructing a scene context library according to the user tag library and the content feature vector library to obtain a scene context library. The present invention solves the problem of lack of intelligent recommendation and prediction in traditional methods by taking users as the center, actively perceiving user needs, and providing accurate and personalized information push and search assistance, thereby improving the intelligence level, user experience and work efficiency of the intelligent workbench.

The application provides a trusted recommendation model based on a double-dimension constraint large language model, and comprises the following steps: S1, a positive and negative sequence construction module: the positive and negative sequence construction module extracts high-score and low-score item sequences from user historical interactions as evidence bases for argumentation; S2, a user preference module: the user preference module counts the historical score distribution of a user to capture the individualized score scale as calibration information for subsequent prediction; S3, a multi-agent argumentation module: the multi-agent argumentation module first generates initial arguments, then mines contradictory evidence through cross-interrogation to form a complete round of argumentation, and a score prediction agent integrates the argumentation result and the user score preference to output the predicted score of the argumentation channel; and S4, a fusion module: the fusion module adjusts the adaptive weight to fuse the statistical result of a traditional recommendation channel and the reasoning result of the multi-agent argumentation channel to generate a final recommendation result, can solve the illusion and position ambiguity problem of a large language model, and enhances the recommendation performance by integrating a deep reasoning large language model.

This invention provides a video recommendation method, apparatus, electronic device, and storage medium. It generates a user status document based on multi-source heterogeneous user information, and then calls a large language model to generate a first semantic unit based on the user status document. Videos matching the first semantic unit are searched from a vector database, and at least a portion of these videos are recommended to the user, thereby improving the video recommendation effect.

The invention discloses a recommendation system optimization method based on adaptive negative sampling and an enhanced graph adjacency matrix, and relates to the technical field of graph recommendation. According to the method, negative samples with proper hardness are adaptively selected in the training process, positive samples with lower rankings and negative samples with higher rankings are more concerned to reduce the influence of FPP and FNP, then a new graph adjacency matrix is introduced, the matrix combines user-user and item-item correlation, and the method has the advantages that the user-user correlation and the item-item correlation are combined, so that the user-user correlation and the item-item correlation are combined, and the user-user correlation and the item-item correlation are combined. A user-project interaction matrix for balancing the interaction number among all users is designed, in order to achieve the purpose, a graph-based recommendation method is pre-trained to obtain embedding of the users and projects, and then the user-project interaction matrix is enhanced through top-K sampling; and symmetric user-user and project-project related components are expanded to an adjacent matrix, so that de-noising is carried out on active users and enhancement is carried out on inactive users.

The application discloses a review-aware recommendation method considering user preference uncertainty, comprising the following steps: 1. using user reviews and ratings to build a data set, 2. building an asymmetric user preference and product feature representation learning model, and fusing user preference and product features to build a rating prediction model, 3. model training and optimization. The application proposes an asymmetric deep learning recommendation method, which comprehensively considers the difference of user reviews in learning user preference and product representation, and introduces user preference uncertainty, so as to improve the product recommendation performance.

The present invention discloses a method and device for mining recommendation information, and relates to the field of computer technology. A specific implementation of the method includes: extracting attribute elements of a set attribute type from text description information of a target object to obtain scene attribute information of the target object; combining the attribute elements of the scene attribute information according to a set combination strategy to generate scene combination information including event attribute elements; generating recommendation information of the target object according to the attribute elements of the scene combination information and a set recommendation template. This implementation combines the attributes of the target object itself with the user's intention, which can not only highlight the target object, but also meet the user's needs, improve the recommendation effect, and enhance the user experience.