Abstract
Recommender systems are information software that retrieves relevant items for users from massive sources of data. The variational autoencoder (VAE) has proven to be a promising approach for recommendation systems, as it can explore high-level user-item relations and extract contingencies from the input effectively. However, the previous variants of VAE have so far seen limited application to domain-specific recommendations that require additional side information. Hence, The Ensemble Variational Autoencoder framework for recommendations (EnsVAE) is proposed. This architecture specifies a procedure to transform sub-recommenders' predicted utility matrix into interest probabilities that allow the VAE to represent the variation in their aggregation. To evaluate the performance of EnsVAE, an instance - called the “Ensemblist GRU/GLOVE model” - is developed. It is based on two innovative recommender systems: 1-) a new “GloVe content-based filtering recommender” (GloVe-CBF) that exploits the strengths of embedding-based representations and stacking ensemble learning techniques to extract features from the item-based side information. 2-) a variant of neural collaborative filtering recommender, named “Gate Recurrent Unit-based Matrix Factorization recommender” (GRU-MF). It models a high level of non-linearities and exhibits interactions between users and items in latent embeddings, reducing user biases towards items that are rated frequently by users. The developed instance speeds up the reconstruction of the utility matrix with increased accuracy. Additionally, it can switch between one of its sub-recommenders according to the context of their use. Our findings reveal that EnsVAE instances retain as much information as possible during the reconstruction of the utility matrix. Furthermore, the trained VAE's generative trait tackles the cold-start problem by accurately estimating the interest probabilities of newly-introduced users and resources. The empirical study on real-world datasets proves that EnsVAE significantly outperforms the state-of-the-art methods in terms of recommendation performances.
Highlights
Recommender systems are software tools conceived to assist users in reaching the most relevant elements from a gigantic collection of resources
Based on the results shown in Fig. 9.b) and Fig. 10.b), one can observe that the Global Vectors for Word Representation (GloVe)-Content-based Filtering models (CBF) recommender obtains a high normalized discounted cumulative gain (NDCG) score on Amazon Electronics dataset and Gated Recurrent Unit layer (GRU)-Matrix Factorization (MF) records a high NDCG score on the MovieLens dataset
In this article, an Ensemblist Variational Autoencoder framework is developed for recommender systems (EnsVAE)
Summary
Recommender systems are software tools conceived to assist users in reaching the most relevant elements from a gigantic collection of resources. Wu et al [46] designed a similar collaborative model based on denoising autoencoders that learn latent representations of corrupted user-item preferences It allows the reconstruction of the rating matrix with a smaller error margin. The hybrid is composed of a set of recommender systems (sub-recommenders), with their rating matrices values adjusted to output interest probabilities Afterward, these matrices are combined using a probabilistic aggregation function and passed to the variational autoencoder that learns the statistical distribution of interests across possible user-item interaction patterns. The proposed model merges a GRU-based Matrix Factorization Recommender System, a neural matrix factorization algorithm reinforced with embeddings adjusted using Gated Recurrent Units It adjusts embeddings for more accuracy reconstruction, with a stacking GloVe- Content-Based Filtring Recommender system, which uses item descriptions and pre-trained GloVe vectors [25] to learn embeddings for each item.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.