Year: 2017
Type: article
Abstract: A recommendation system is a program which attempts to predict items that users may be interested in, considering their preference and taste.Collaborative filtering is an algorithm which is most commonly used in the recommendation system.It recommends the items that users are interested in with similar preferences.The system finds similar users based on all users' purchase and rating data.However, user preferences can change, and this kind of change may slow down recommendation performances.To upgrade these recommendation performances, this study proposes a user preference-reflected recommendation technique.The precision and MAE of the proposed method were measured, using MovieLens data set.Compared to conventional collaborative filtering techniques, the proposed method revealed better results.
Author: Sumi Shin
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W2569946699
