A Review of Text-Based Recommendation Systems

Many websites over the Internet are producing a variety of textual data; such as news, research articles, ebooks, personal blogs, and user reviews. In these websites, data is so large that process finding pertinent information by often becomes cumbersome. To overcome this issue, “Text-based Recommendation Systems (RS)” being developed. They systems with capability to find relevant in minimal time using text primary feature. There exist several techniques build evaluate systems. And though good number surveys compile general attributes recommendation systems, there still lack comprehensive literature review about text-based paper, we present latest studies on RS. We have conducted survey collecting from preeminent digital repositories, was published during period 2010-2020. This mainly covers four major aspects based used reviewed literature. The datasets, feature extraction techniques, computational approaches, evaluation metrics. As benchmark datasets carry vital role any research, publicly available extensively paper. Moreover, for RS many proprietary also used, which not public. But consolidated all publically familiarize new researchers. Furthermore, methods briefed their usage construction discussed. Later, various approaches use features some metrics adopted. presented an overview diagramed them according popularity. concludes Word Embedding widely selection technique research. deduces hybridization other enhance accuracy. study highlights fact most work English data, News popular domain.