Recommendation system with multi-dimensional and parallel-case four-term analogy

Recommendation systems on the Internet have become more necessary due to enormous amounts of information that keep increasing. Existing recommendation systems, such as Content-Based Filtering (CBF) and Collaborative Filtering (CF), have a trade off: recommended items cannot reflect users´ preferences and offer valid unexpected elements at the same time. Our goal is to resolve this trade-off problem. We propose a recommendation system that uses four-term analogy, which is a way of thinking. We prove the proposed system´s effectiveness by comparing it with existing systems.