Context-Aware Approach for Restaurant Recommender Systems

Context-Aware Approach for Restaurant Recommender Systems

Haoxian Feng (University of Ottawa, Canada) and Thomas Tran (University of Ottawa, Canada)
Copyright: © 2018 |Pages: 15
DOI: 10.4018/978-1-5225-2255-3.ch153


This paper addresses the issue of how to effectively use users' historical data in restaurant recommender systems, as opposed to systems, such as FindMe, that only rely on online operations. Towards that end, the authors propose a bias-based SVD method as the underlying recommendation algorithm and test it against the traditional item-based collaborative filtering method on the Entrée restaurant dataset. The results are promising as the obtained Root-Mean-Square-Error (RMSE) values reach 0.58 for the SVD and 0.62 for the item-based system. Researchers can extend the transformation from user behaviors to ratings in more application domains other than the restaurant one.
Chapter Preview

Upsurge of Recommender Systems

In the 90s of the last century, the field of electronic commerce began to flourish and personalized recommender system also rose its first wave. Some large e-commerce sites such as Amazon launched their own personalized recommender system. Literature reported that 35 percent of Amazon's incremental sales came from their recommender system (Figiel, Epstein, McDonald, 1998, pp.20-25). At the same time, a sensational paper published by Amazon in 2000, entitled “Item-based collaborative filtering recommendation algorithms” (Sarwar, et al., 2001), has become one of the most famous literature in this area. Ever since then, personalized recommendation technology has become an irresistible trend.

Key Terms in this Chapter

Baseline Predictors: A method that considers the user bias, the item bias and the global average of ratings in the dataset to provide better recommendation results.

Context-Aware Recommender System: A recommender system that considers not only the ratings but also includes other information such as time, emotion and so on.

Gradient Descent: A fundamental optimization algorithm that is used to find the steepest descent direction and the local minimum.

Item-Based Collaborative Filtering: A classical and most widely used recommendation algorithm that uses the similarities between items.

Singular Value Decomposition (SVD): One of the matrix factorization methods to decompose a matrix to the product of two orthogonal matrices and one diagonal matrix.

Matrix Factorization: A way to decompose a matrix to the product of several matrices.

Behavior-Score Mapping: A transformation process from the behavior to the ratings under some rules.

Complete Chapter List

Search this Book: