On the Smaller Number of Inputs for Determining User Preferences in Recommender Systems

• DOI: 10.3390/math8122138
• Publication Date: 2020-12-01T18:08:51Z
• Authors: Sang-Min Choi; Dongwoo Lee; Chihyun Park
• Keywords: Ask price; User needs; User requirements document

One of the most popular applications for recommender systems is a movie recommendation system that suggests few movies to user based on user’s preferences. Although there wealth available data movies, such as their genres, directors and actors, little information new user, making it hard suggest what might interest user. Accordingly, several services explicitly ask users evaluate certain number which are then used create profile in system. In general, one can better if evaluates many at beginning. However, do not want when they join service. This motivates us examine minimum inputs needed reliable preference. We call this magic determining A reduce inconvenience while also suggestions. Based item content-based filtering, we calculate by comparing accuracy resulting from use different numbers predicting