Critical Analysis of Various Recommendation Systems

  • Muzamil Shah
  • Fazli Subhan


Recommender systems predict the requirements and interests of users-based on historical data.
All the users having same liking for an item are clustered in the same group through which the recommender
system identifies the strengths or pertinent likelihood of an item/product. Different techniques are used to
predict the user preferences such as content-based system, collaborative-based filtering, knowledge-based
system, multi-attribute trader and e-commerce. Among the several proposed techniques presented in the
literature, the content-based and collaborative approaches are the most efficient and effective techniques.
These techniques are used in all the hybrid techniques. The content-based identify the description of items
and collaborative techniques collect the same preference of users in the same cluster or group. In this paper,
we present a survey of recommender systems proposed in the contemporary literature. We have studied the
various techniques being used in the recommendation systems and have evaluated them critically by finding
their strengths and weaknesses. We have also identified the research gaps related to these techniques and
have also suggested the probable improvements in the proposed techniques to make them more effective.
The critical analysis of the recommendation techniques is presented in this paper.