Neural embedding collaborative filtering for recommender systems
- PDF / 1,103,104 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 35 Downloads / 271 Views
(0123456789().,-volV)(0123456789(). ,- volV)
ORIGINAL ARTICLE
Neural embedding collaborative filtering for recommender systems Tianlin Huang1 • Defu Zhang1 • Lvqing Bi1 Received: 4 July 2019 / Accepted: 6 April 2020 Ó Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract The main purpose of collaborative filtering algorithm is to provide a personalized recommender system based on past interactions of each user (e.g., clicks and purchases). Among various collaborative filtering techniques, matrix factorization is widely adopted in diverse applications. This technique has superior characteristics, including applying latent feature vectors to represent users or items and projecting users and items into a shared latent feature space. In the present study, a matrix factorization model with the neural embedding called the neural embedding collaborative filtering (NECF) is proposed. In order to evaluate the performance of the proposed method, a probabilistic auto-encoder is initially applied to achieve unsupervised learning to generate the neural embedding vector from the user–item data. Secondly, these vectors are combined with a regression model based on single point negative sampling to represent the latent feature vectors of the user with regression coefficients. Moreover, an inner product is applied on latent features of users and items to determine the correlations between them. It should be indicated that the NECF is generic so that it can express and generalize the matrix factorization under its framework. In the present study, a ridge regression learning is applied on latent features of each user. The experimental results on two benchmark data sets show that the proposed model outperforms other state-ofthe-art methods. Keywords Collaborative filtering Neural embedding Matrix factorization Implicit feedback
1 Introduction In the era of the information explosion, information overload is an enormous challenge that we are confronted with. Meanwhile, personalized recommendations enhance the user experience and expand product margins. The key to personalized recommender systems is to model preferences of the user for different aspects in accordance with the past interactions of each user. The primary method in this regard is the collaborative filtering. Reviewing the literature indicates that several methods have been proposed so far for the collaborative filtering. These methods can be
& Tianlin Huang [email protected] Defu Zhang [email protected] Lvqing Bi [email protected] 1
School of Information Science and Engineering, Xiamen University, Xiamen 361005, China
mainly divided into three categories, including memorybased or neighborhood-based approaches, latent factor models (LFM) and hybrid models [12]. With the promotion of the Netflix awards, matrix factorization (MF) algorithm has become a practical approach for recommendation system based on underlying factor models. In this approach, each user and item is projected into a common low-dimensional space, where latent feature ve
Data Loading...
