A survey on recommender system techniques and their applications
Recommender systems are web-based systems which help in the reduction or eradication of information overload in an information system. Recommender systems study the characteristics of a system user by identifying their ratings, purchases or other demographic attributes and then use the information gathered on the user to subsequently provide recommendation of items to the user. The design of the recommender system has to do with the use of various techniques such as the collaborative filtering technique, the content-based technique, and the hybrid technique. The collaborative filtering technique involves knowing the similarity between users and how they correlate with each other as a result of their activities on the platform. The content based filtering involves the identification of the contents and attributes of various items provided to the users while the hybrid technique involves the combination of both the collaborative and content based filtering techniques. The application of any of these techniques depends on the dataset available, area of application and also the performance expected of the design. This paper provides an extensive study on these techniques, stating their applications, their advantages and disadvantages. The outcome presented researchers with choices while implementing recommender systems.