Improving diversity using bandwagon effect for developing recommendation system

The recommendation system using collaborative filtering (CF) methods is widely used. However, it is short of recommending only similar items that are popular with users. To break this limitation of CF method, we design the recommending system based on the psychology concept of bandwagon effect. Generally, consumers decide what they are going to purchase based on what others have purchased. This is called bandwagon effect. To design the recommendation system based on bandwagon effect, we use the matrix factorization (MF) based alternating least square (ALS). Moreover, to store big data and computing, we construct a cluster based on in-memory framework spark and accomplish the development and computing of recommendation system. In order for improving the recommendation diversity, we compare the recommendation list from the existing recommendation system and our proposed recommendation system and it showed that our proposed system indicated better diversity during recommendation.