A Deep Hybrid Model for Recommendation Systems

TL;DR

This paper introduces a deep hybrid neural network model for recommendation systems that leverages both ID embeddings and auxiliary features, demonstrating improved performance on a job recommendation dataset.

Contribution

The paper presents a novel deep neural network architecture that integrates auxiliary features with ID embeddings for enhanced recommendation accuracy.

Findings

Improved recommendation accuracy over models using only ID embeddings.

Effective utilization of categorical and continuous features in deep learning models.

Demonstrated success on a real-world job-site dataset.

Abstract

Recommendation has been a long-standing problem in many areas ranging from e-commerce to social websites. Most current studies focus only on traditional approaches such as content-based or collaborative filtering while there are relatively fewer studies in hybrid recommender systems. Due to the latest advances of deep learning achieved in different fields including computer vision and natural language processing, deep learning has also gained much attention in Recommendation Systems. There are several studies that utilize ID embeddings of users and items to implement collaborative filtering with deep neural networks. However, such studies do not take advantage of other categorical or continuous features of inputs. In this paper, we propose a new deep neural network architecture which consists of not only ID embeddings but also auxiliary information such as features of job postings and candidates for job recommendation system which is a reciprocal recommendation system. Experimental results on the dataset from a job-site show that the proposed method improves recommendation results over deep learning models utilizing ID embeddings.