SOCIAL MEDIA-BASED RECOMMENDER SYSTEMS: A REVIEW OF ARCHITECTURES AND LIMITATIONS

Recommender systems are tools for relating with huge and seemly complex information spaces. Good recommendation system should provide potential customers with the most relevant and applicable products regardless of information overload. Social media has become a huge influence to product recommendation owing to robust sets of data made available to recommendation systems. The synergy between social media data and recommendation systems is therefore of outmost importance and calls for a need to conduct an in-depth survey of several emerging social-media based recommender systems. Investigating recommendation systems based on dataset from social networks such as dataset from social media such Netflix, Facebook, Twitter, Epinions, Flixster, Amazon, Sina Weibo, Jingdong, among few others, collaborative filtering was realized to be the main approach of many authors. These authors opted for collaborative filtering based due to its robustness in handling coldstart problems. This study thence improved the already exiting social media based recommendation system body of knowledge by unifying frequent social and buying human behaviour, and recommendation systems