Hybrid Insurance Recommendation Algorithm Integrating Deep Neural Networks and Knowledge Graphs Based on Matrix Factorization

In multifactorial systems using textual and graphical information in matrix factorization to facilitate the problem of separate data processing. Recently, in some studies, the study of neural networks to understand the content of text and graphic elements more deeply and to achieve efficacy by creating more accurate patterns of recognition of elements. However, the open question remains about how to effectively use graphic data from the thermal imager in matrix factorization. In this paper, we proposed a double-regularized matrix factorization with deep neural networks (DRMF) to solve this problem. DRMF applies a multilayered neural network model by stacking a convolutional neural network and a secured repetitive neural network to create independent distributed views of user content and objects. Then representations serve to regularize the generation of hidden models for both users and for elements of matrix factorization. So the proposed new model of the neural network works better than a model with a single convergent neural network. In this paper, we propose double - regularized matrix factorization with deep neural networks (DRMF) to solve this problem. DRMF uses a multi-layered neural network model by enclosing a convoluted neural network and a secure repeating neural network to create independent distributed representations of user content and objects. Then the representations are used to regularize the generation of hidden models for both users and elements of matrix factorization. Thus, the proposed new neural network model works better than the model with a single converging neural network. In traditional SF methods, only a feedback matrix is ​​used, which contains explicit (eg, estimates) or implicit feedback to train and predict the life of the motor. As a rule, the feedback matrix is ​​liquid, which means that most users encounter several elements. Based on this was presented in Proc. BigData Congress. However, this view has been significantly expanded using a new deep neural network model and adding new experimental attachments compared to the conference publication.

Dual-regularized matrix factorization with deep neural networks for recommender systems

A multi-attention deep neural network model base on embedding and matrix factorization for recommendation

All current recommendation algorithms, when modeling user–item interactions, basically use dot product. This dot product calculation is derived from matrix factorization. We argue that an inherent drawback of matrix factorization is that latent semantic vectors of users or items sometimes do not satisfy triangular inequalities, which may affect the performance of the recommendation. Recently, metric factorization was proposed to replace matrix factorization and has achieved some improvements in terms of recommendation accuracy. However, similar to matrix factorization, metric factorization still uses a simple, linear fashion. In this paper, we explore leveraging nonlinear deep neural networks to realize Euclidean distance interaction between users and items. We propose a generic Neural Metric Factorization Framework (NMetricF), which learns representations for users and items by incorporating Euclidean metric factorization into deep neural networks. Extensive experiments on six real-world datasets show that, compared to the previous recommendation algorithms based purely on rating data, NMetricF achieves the best performance.

The rapid development of the Internet of Things (IoT) and e-commerce has brought a lot of convenience to people’s lives. IoT applications generate a large number of services and user data. It is necessary to design a personalized recommendation technology suitable for the users of IoT services and improve the user experience. In this paper, a recommendation algorithm with trusted relevance combined with matrix factorization is proposed. By establishing an effective trust metric model, the user’s social information is integrated into the recommendation algorithm. First, the social network concentric hierarchical model is used to consider the direct or indirect trust relationship, and more trust information is integrated for the matrix factorization recommendation algorithm. Then, we design the trust relevance, comprehensively considering the trust factors and interest similar factors. Our experiments were performed on the Dianping datasets. The recommendation algorithm using matrix factorization and trusted relevance degree has higher prediction accuracy than the basic matrix decomposition and social matrix factorization in terms of accuracy and stability.

Recommendation systems are widely used in various industries and are seen as one of the effective methods in reducing information overload. This paper selected four common algorithms for implementation and evaluation among many recommendation algorithms. Two traditional collaborative filtering algorithms - User Collaboration Filter and Item-based Collaborative Filtering; most popular algorithms in the recommendation field - Matrix Factorization; and Neural Collaborative Filtering - the algorithm based on Matrix Factorization and combined with neural networks.

Collaborative filtering is the most classical technology in recommendation system. Compared with memory-based collaborative filtering technology, matrix factorization has good scalability and recommendation effect, which makes it widely used. On the basis of matrix factorization model, deep neural network is introduced to improve the accuracy of scoring prediction and the quality of recommendation. Experiments on MovieLens dataset show that the proposed method improves the accuracy and quality of recommendation algorithm.

INTRODUCTION: Collaborative filtering (CF) algorithms based on deep learning and matrix factorization aim to learn deeper latent features of users and items from user ratings. However, such methods often face challenges due to the sparsity of rating data, which limits their recommendation performance. OBJECTIVES: This study aims to design a hybrid recommendation algorithm named auto-encoder deep learning and matrix factorization (AED-MF), which integrates deep learning and matrix factorization to address the sparsity issue in rating data and improve the extraction of deeper latent features of users and items. METHODS: The AED-MF algorithm combines an auto-encoder-based deep learning approach with matrix factorization to model both deeper hidden representations and nonlinear relationships between users and items. The methodology includes data downloading, mounting, cleaning, model training, and experimental evaluation, with matrix factorization applied to predict missing ratings in the rating matrix. RESULTS: The proposed AED-MF algorithm demonstrated strong performance in key recommendation metrics, effectively learning deeper user and item representations and applying them to capture complex nonlinear relationships. It also reduced the impact of rating data sparsity while maintaining high recommendation accuracy. CONCLUSION: The AED-MF recommendation algorithm successfully alleviates the problem of sparse rating data and enhances recommendation accuracy by leveraging the strengths of both deep learning and matrix factorization, offering an effective solution for improving collaborative filtering-based recommender systems.

Recognizing consumer interests and satisfaction is crucial for organizations to thrive in a competitive market.A strong marketing strategy aids companies in reaching customers and building a reputation for their products.Presently, numerous internet platforms employ recommender systems on a commercial basis.Customer preferences are generally obtained behaviours of users from online shopping which is used to recommend the products.Deep Learning (DL) models have been used for efficient customer preference recommendation, but these models don't consider secondary information such as Knowledge Graph (KG) and comment information.Hence, in this paper, Time Aware Attention and KG embedded Deep Network for Recommendation (TAK-DepNR) is proposed by combining Convolutional Neural Network (CNN) and Graph Convolutional Network (GCN).In this method, the user similarity, item similarity and latent period are modelled using Matrix Factorization (MF) model.The MF generates long and short preferences which are then fed into the CNN and GCN models to produce dynamic user preferences.The CNN extracts higher-level characteristics among the user and item from MF, then reduces to lower-order features.The GCN aids to train the interpretation of each node and encodes the KG based Global Convolutional Feature Map (KGGCFM) for CNN through the shared features.This creates a major link between the products and allowing for more extensive item feature description for user preference graph.The CNN will be integrated with GCN using its squeezed units to learn the high-level feature relations between recommendation systems and KG entities.Finally, dynamic user preferences from CNN and GCN are passed into Fully Connected (FC) layer constructed network to recommend the items.The experimental results show that the TAK-DepNR realizes average precision of 89% and 88% compared to the other models like Immersive Graph Neural Network (IGNN), Attentional Factorization Machine with Review-Based User-Item Interaction (AFMRUI), Item Collaborative Filtering with RNN (ICFRNN), Deep Reinforcement Recommender System with Maximum pooling layers (DRR-Max) and Sentiment Analysis with Bidirectional Long Short Term Memory (SA-BiLSTM) on Amazon Digital Music Dataset and Book Crossing Dataset.

Matrix factorisation is a one of the most popular techniques in recommendation systems. However, matrix factorisation still suffers from cold start problem and needs complicated computation. In this paper, we present a hybrid recommendation algorithm, which integrates user and item content information and matrix factorisation. First, based on user or item content information, biases of user or item can be evaluated in advance. Incorporating user and item biases into matrix factorisation model, we can obtain final prediction model. At last, momentum stochastic gradient descent method is used to optimise other parameters. Experimental results on a real data set have shown best performance of our algorithm in terms of MAE and RMSE when compared with other classical matrix factorisation recommendation algorithms.

In the paper, the influence of user trust relationship on recommendation is mainly analyzed, user rating and trust relationship are also clearly defined.Meanwhile, the traditional SVD matrix factorization algorithm is extended to introduce the explicit influence (user trust value) and the implicit influence (friends trusted by users) of the user trust into the matrix factorization recommendation model.Subsequently, TFMF (Trust-fused Matrix Factorization) recommendation model is proposed in the paper to infer model training & learning in detail.The user trust relationship is not considered in implicit feedback SVD++, in other words, the itemratings marked by the trusted users can also influence the recommendation prediction.Therefore, the matrix factorization recommendation algorithm based on implicit trust relationship fusion is designed in the paper.Compared with original matrix factorization algorithm, the proposed algorithm not only has better prediction accuracy and flexibility, but also has better performance in the sparse data set due to the introduction of the trust relationship implicit feedback mechanism.

Matrix factorisation is a one of the most popular techniques in recommendation systems. However, matrix factorisation still suffers from cold start problem and needs complicated computation. In this paper, we present a hybrid recommendation algorithm, which integrates user and item content information and matrix factorisation. First, based on user or item content information, biases of user or item can be evaluated in advance. Incorporating user and item biases into matrix factorisation model, we can obtain final prediction model. At last, momentum stochastic gradient descent method is used to optimise other parameters. Experimental results on a real data set have shown best performance of our algorithm in terms of MAE and RMSE when compared with other classical matrix factorisation recommendation algorithms.

A topic sentiment based method for friend recommendation in online social networks via matrix factorization

Recommendation system has been placed much emphasis by researchers and programmers to deal with the information overload. Collaborative filtering algorithm is the most commonly used one. In order to enhance its performance, the Matrix Factorization was discovered to base the collaborative filtering. This paper elaborates on the collaborative filtering algorithm based on Matrix Factorization and gives a brief introduction of the gradient descend technique adopted by Matrix Factorization. Then the paper shows the whole procedures of Basic MF (Matrix Factorization), Regularized MF and Biases MF. The experiment result of collaborative recommendation algorithms based on three types of MF have been analyzed and compared, helping determine the evaluation of the number of latent factors K, the time of iteration K and regularization coefficient λ and optimizing the accuracy of algorithms.

Learning matrix factorization with scalable distance metric and regularizer