Recommendation System for videos on the Internet

点击翻译按钮获取中文摘要

In this research we proposed and analyzed a choice model that utilizes two types of data, behavioral and(stated) preference data, in order to improve the accuracy of a recommendation system. We applied the model to videos viewing and preferences data of a particular video-sharing website on the internet. We found that compared with the model that used only behavioral and only preference data, the model that combined both data explained the variation of data and predicted individual decisions much better. This indicates the value of using both data in recommendation system.