Recommendation Enhancement Using Traceability and Machine Learning: A Review

Abstract

For many years now, the available data on the internet became more and more complex, and its volume is increasing at a terrifying speed. Nowadays, filtering these data and selecting the ones which are the most suitable for our needs is the real challenge. Many researchers have proposed intelligent filtering systems that proved their efficiency and managed to recommend only a selection of items, able to meet the initial expectations. Most of these systems employ both collaborative and content-based techniques but within different approaches, often combined to provide the most suitable results. As input, they exploit textual and history data from users’ activities, taken either from their profiles or from other valuable resources. However, due to the complexity of data, the recommendation becomes more redundant, and the same elements get proposed repeatedly. In this paper, we present a literature review of traceability and interesting recommendation approaches, and we intend to discuss the benefits of using traceability and machine learning and how they can improve the recommendation and make it more efficient and reliable.