Tatari: An open source software tool for the development and evaluation of recommender system algorithms

Recommender systems give recommendations on items to users. They are widely used in e-commerce systems to locate products that could be of interest to users. Over the years various systems have been developed with researchers developing new algorithms or enhancing existing ones. This paper describes Tatari(1), an open source software tool developed at the University of Auckland that enables researchers to develop and test recommender algorithms and compare them against others to analyze results. Tatari can also be used as an online recommender system to help users locate items of interest. In this paper we outline the motivation, design, architecture, implementation and operation of Tatari. The paper concludes by describing possible future developments.