Abstract
Matrix Factorization (MF) is one of the most successful Collaborative Filtering (CF) techniques used in recommender systems due to its effectiveness and ability to deal with very large user-item rating matrix. However, when the rating matrix sparseness increases its performance deteriorates. Expanding MF to include side-information of users and items has been shown by many researchers both to improve general recommendation performance and to help alleviate the data-sparsity and cold-start issues in CF. In regard to item feature side-information, most schemes incorporate this information through a two stage process: intermediate results (e.g., on item similarity) are first computed based on item attributes; these are then combined with MF. In this paper, focussing on item side-information, we propose a model that directly incorporates item features into the MF framework in a single step process. The model, which we name FeatureMF, does this by projecting every available attribute datum in each of the item features into the same latent factor space with users and items, thereby in effect enriching item representation in MF. Results are presented of comparative performance experiments of the model against three state-of-the-art item information enriched models, as well as against four reference benchmark models, using two public real-world datasets, Douban and Yelp, with four training:test ratio scenarios each. It is shown to yield the best recommendation performance over all these models across all contexts including data-sparsity situations, in particular, achieving over 0.9% to over 6.5% MAE recommendation performance improvement over the next best model, HERec. FeatureMF is also found to alleviate cold start and to scale well, almost linearly, in regard to computational time, as a function of dataset size.
Highlights
In the modern ‘Big Data’ era, recommender systems have become the essential tools that address a myriad of services for Internet users, e.g., in the context of the vast range of information and services readily available such as web browsing [1], [2] or IoT scenarios [3], by assisting users to discover what they need or receive timely personalized suggestions and recommendations
The results indicate that FeatureMF is able to utilize item features more effectively to make better recommendations
A novel Matrix Factorization (MF) model, FeatureMF, has been proposed and set out in this paper. It exploits item features information where available to improve recommendation performance. This novelty primarily lies in the way it diffuses the item latent factors, derived for example from global ratings, with latent factor representations of attributes of item features into a single MF computational stage
Summary
Several works report modeling item features as a HIN, and incorporate results from a HIN into MF to address the data-sparsity and cold-start problems, and yield improved performances on [16] and [17]. Some of the similarity measures have limitations for features, for example, similarity measures proposed in Nguyen and Zhu et al [16] and Yu et al [17], can only deal with the situation that, for each item, every feature can only contain one attribute value This characteristic of the scheme constrains its use in performance comparison experiments such as those reported on here below. The novel item feature enriched MF model proposed incorporates item information directly into the MF framework, without a pre-processing stage, and allows for one or more attributes of each item feature.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
