Abstract:Abstract The sparsity problem remains a significant bottleneck for recommendation systems. In recent years, deep matrix factorization has shown promising results in mitigating this issue. Furthermore, many works have improved the prediction accuracy of deep matrix factorization by incorporating the user’s and/or items’ auxiliary information. However, there are still two remaining drawbacks that need to be addressed. First, the initialization of latent feature representations has a substantial impact on the performance of deep matrix factorization, and most current models utilize a uniform approach to this initialization, constraining the model’s optimization potential. Secondly, many existing recommendation models lack versatility and efficiency in transferring auxiliary information from users or items to expand the feature space. This paper proposes a novel model to address the issues mentioned above. By using a semi-autoencoder, the pre-trained initialization of the latent feature representation is realized in this paper. Simultaneously, this model assimilates auxiliary information, like item attributes or rating matrices from diverse domains, to generate their latent feature representations. These representations are then transferred to the target task through subspace projection distance. With this, this model can utilize auxiliary information from various sources more efficiently and this model has better versatility. This is called deep matrix factorization via feature subspace transfer. Numerical experiments on several real-world data show the improvement of this method compared with state-of-the-art methods of introducing auxiliary information about items. Compared with the deep matrix factorization model, the proposed model can achieve 6.5% improvement at most in the mean absolute error and root mean square error.

Multi-Feature Extension Via Semi-Autoencoder for Personalized Recommendation

Exploiting SDAE Model for Recommendations

Representation Learning: Serial-Autoencoder for Personalized Recommendation

Personalized recommendation with knowledge graph via dual-autoencoder

Representation Learning: Recommendation With Knowledge Graph via Triple-Autoencoder

Hybrid Collaborative Recommendation via Semi-AutoEncoder

Self-Attention Mechanism Enhanced User Interests Modeling for Personalized Recommendation Services in Cyber-Physical-Social Systems.

A Hybrid Recommender System Based on Autoencoder and Latent Feature Analysis

Representation Learning with Collaborative Autoencoder for Personalized Recommendation

Attentive Autoencoder Matrix Factorization for Recommender Systems

Hybrid Collaborative Recommendation via Dual-Autoencoder

Hybrid collaborative recommendation of co-embedded item attributes and graph features

Point-of-Interest Recommendation: Exploiting Self-Attentive Autoencoders with Neighbor-Aware Influence

Hybrid Collaborative Filtering with Semi-Stacked Denoising Autoencoders for Recommendation

Research of Personalized Recommendation System Based on Multi-view Deep Neural Networks

Recommendation System Using Autoencoders

Deep matrix factorization via feature subspace transfer for recommendation system

An improved hybrid semi-stacked autoencoder for item-features of recommendation system (iHSARS)

SMEF: Social-aware Multi-dimensional Edge Features-based Graph Representation Learning for Recommendation

Knowledge-Aware Multispace Embedding Learning for Personalized Recommendation

Large Multi-modal Encoders for Recommendation