Date of Award

6-2018

Document Type

Dissertation

Publisher

Santa Clara : Santa Clara University, 2018.

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Engineering

First Advisor

Yi Fang

Abstract

The mass adoption of the internet has resulted in the exponential growth of products and services on the world wide web. An individual consumer, faced with this data deluge, is expected to make reasonable choices saving time and money. Organizations are facing increased competition, and they are looking for innovative ways to increase revenue and customer loyalty. A business wants to target the right product or service to an individual consumer, and this drives personalized recommendation. Recommender systems, designed to provide personalized recommendations, initially focused only on the user-item interaction. However, these systems evolved to provide a context-aware recommendations. Context-aware recommender systems utilize additional context, such as genre for movie recommendation, while recommending items to users. Latent factor methods have been a popular choice for recommender systems. With the resurgence of neural networks, there has also been a trend towards applying deep learning methods to recommender systems.

This study proposes a novel contextual latent factor model that is capable of utilizing the context from a dual-perspective of both users and items. The proposed model, known as the Group-Aware Latent Factor Model (GLFM), is applied to the event recommendation task. The GLFM model is extensible, and it allows other contextual attributes to be easily be incorporated into the model. While latent-factor models have been extremely popular for recommender systems, they are unable to model the complex non-linear user-item relationships. This has resulted in the interest in applying deep learning methods to recommender systems. This study also proposes another novel method based on the denoising autoencoder architecture, which is referred to as the Attentive Contextual Denoising Autoencoder (ACDA). The ACDA model augments the basic denoising autoencoder with a context-driven attention mechanism to provide personalized recommendation. The ACDA model is applied to the event and movie recommendation tasks.

The effectiveness of the proposed models is demonstrated against real-world datasets from Meetup and Movielens, and the results are compared against the current state-of-the-art baseline methods.

Share

COinS