Detection and Correction of Natural Noise in the Recommendation System

Recommender systems have become a mainstream research field in IT. Recommender systems act as a supporting tool for helping users trying to obtain information that best fits their preferences and needs, in scenarios where the information overload is an important drawback. The most popular version of RS (recommender systems), learns from preferences of the user about a predefined set of known items, and predicts the preference degree about unknown items. With this goal, many applications have been built to recommend different types of items like music, books, TV shows, movies, jokes, news, scientific papers, web pages, and so on. They have covered diverse areas like e-commerce, e-learning, e-services, tourism, and software engineering. The two most used techniques in the development of recommender systems are the content-based and the collaborative filtering. Content-based recommender systems, suggest items with similar features to those that the user chose in the past. On the other hand, collaborative filtering recommender systems recommend items that other similar users liked in the past.