Search

Traditional collaborative filtering (CF) recommendation algorithms usually use a single rating to recommend items to users, which works well in terms of predictive accuracy. However, recent research on multi-criteria recommender has shown that multi-criteria ratings are of great value to improving recommendation performance. In this paper, we present novel multi-criteria recommendation schemes which leverage multi-criteria ratings and codebook cluster information derived from user-item-criteria ratings matrix to enhance recommendation quality. Particularly, we utilize Factorization Machines (FMs) to integrate the codebook clusters information on individual criteria, which contains users’ preferences on different criteria of items, to extend user-item-criteria interaction feature vectors and make an overall rating prediction. A set of experiments on a real-world datasets show that our approach outperforms both FMs-based single-rating recommendation algorithms in which the clusters information of users or items are based on an overall rating, as well as three existing state-of-the-art multi-criteria recommendation algorithms even in case where data are under high sparsity.

The problem of personalized review-based rating prediction aims at inferring users’ ratings over their unrated items using historical ratings and reviews. Most of existing methods solve this problem by integrating latent factor model and topic model to learn interpretable user and items factors. However, these methods ignore the word order in reviews and the learned topic factors are limited to review text, which cannot fully reveal the complicated interaction relations between reviews and ratings. Moreover, they only utilize overall ratings instead of multi-criteria ratings which can represent more detailed preferences of users. In this paper, we propose a deep learning framework named NRPMT in which multi-criteria ratings and user reviews can complement each other to improve recommendation accuracy. The proposed model can simultaneously predict accurate ratings and generate review content expressing user experience and feelings. For rating prediction, a neural factorization machines-based regression model are used to project the feature interactions between user, item and criteria into rating. For review generation, gated recurrent neural networks are employed to “translate” the feature representation of user, item and criteria into a review. Extensive experiments on three real-world datasets demonstrate that NRPMT achieves significant improvement over the several competitive baselines.

Local Binary Pattern (LBP) has achieved great success in texture classification due to its accuracy and efficiency. Traditional LBP method encodes local features by binarying the difference in local neighborhood and then represents a given image using the histogram of the binary patterns. However, it ignores the directional statistical information. In this paper, some directional statistical features—including the mean and standard deviation of the local absolute difference—are integrated into the feature extraction to improve the classification ability of the extracted features. In order to reduce estimation errors of the local absolute difference, we further utilize the least square estimate technique to optimize the weight and minimize the local absolute difference, which leads to more stable directional features. In addition, a novel rotation invariant texture classification approach is presented. Experimental results on several texture and face datasets show that the proposed approach significantly improves the classification accuracy of the traditional LBP.

With the rapid development of information society, the era of big data is coming. Various recommendation systems are developed to make recommendations by mining useful knowledge from massive data. The big data is often multi-source and heterogeneous, which challenges the recommendation seriously. Collaborative filtering is the widely used recommendation method, but the data sparseness is its major bottleneck. Transfer learning can overcome this problem by transferring the learned knowledge from the auxiliary data to the target data for cross-domain recommendation. Many traditional transfer learning models for cross-domain collaborative recommendation assume that multiple domains share a latent common rating pattern which may lead to the negative transfer, and only apply to the homogeneous feedbacks. To address such problems, we propose a new transfer learning model. We do the collective factorization to rating matrices of the target data and its auxiliary data to transfer the rating information among heterogeneous feedbacks, and get the initial latent factors of users and items, based on which we construct the similarity graphs. Further, we predict the missing ratings by the twin bridge transfer learning of latent factors and similarity graphs. Experiments show that our proposed model outperforms the state-of-the-art models for cross-domain recommendation.