AutoRec: A Recommender System Based on Social Media Stream

In this study, we employed Collaborative Filtering and Sentiment Analysis. This is to automatically produce a recommendation to consumers without requiring them to input ratings and feedback's. The researcher will be using Social media stream as test-bed which was currently being utilized by researchers around the globe to generate information which are useful in different fields. TwitterAPI will be employed to gather real time tweets to be stored in a database for analysis. Framework will be introduced to automatically produce a real time recommendation, which aims to test the effectiveness of the framework proposed by virtue of using DeLone and McClean IS Success Model. As a result the prototype datasets crawled from real world social media stream demonstrates that the proposed method depicts to be constructive. This study can contribute to the growing research projects on recommender system, data mining and to other relevant techniques.