Is Simple Better? Revisiting Non-Linear Matrix Factorization for Learning Incomplete Ratings

Abstract

Matrix factorization techniques have been widely used as a method for collaborative filtering for recommender systems. In recent times, different variants of deep learning algorithms have been explored in this setting to improve the task of making a personalized recommendation with user-item interaction data. The idea that the mapping between the latent user or item factors and the original features is highly nonlinear suggest that classical matrix factorization techniques are no longer sufficient. In this paper, we propose a multilayer nonlinear semi-nonnegative matrix factorization method, with the motivation that user-item interactions can be modeled more accurately using a linear combination of non-linear item features. Firstly, we learn latent factors for representations of users and items from the designed multilayer nonlinear Semi-NMF approach using explicit ratings. Secondly, the architecture built is compared with deep-learning algorithms like Restricted Boltzmann Machine and state-of-the-art Deep Matrix factorization techniques. By using both supervised rate prediction task and unsupervised clustering in latent item space, we demonstrate that our proposed approach achieves better generalization ability in prediction as well as comparable representation ability as deep matrix factorization in the clustering task.

Augmenting organizational decision-making with deep learning algorithms: Principles, promises, and challenges

“…Consumers who interact through digital media leave traces of information as digital footprints in diverse forms, such as text, digital images, login information, and GPS location data. Such digital traces provide managers with considerable opportunities to draw insights using DL and, in turn, target appropriate customer groups with product recommendation (Zheng, Noroozi, & Yu, 2017;Krishna, Guo, & Antulov-Fantulin, 2018) and marketing decisions (Batra & Keller, 2016). A case in point is Unilever partnering with Alibaba to apply DL for spotting customer needs and market development patterns, while optimizing online and offline demand-generation activities.…”

Signed Latent Factors for Spamming Activity Detection

“…Furthermore, numerous collected users may only be related to a few numbers of items/targets in the experimental graph, generating a highly sparse relational matrix, which further aggravates the degree of graph incompleteness. Another motivation to introduce the thought of latent factors to the anti-spam area is evoked, since it is well known that the latent factors of MF are mainly used to learn incomplete or potential interactions [17,49].…”