Over the past year, the size of data available to users on online media has increased exponentially. Due to this large amount of data, people face problems in scaling all this available data as they do not have enough time and hence are unable to find useful items. So, to overcome this type of problem, recommendation system plays an important role. The system filters the data source and provides them with useful information, when this information is in the form of suggestions, the system called recommendation system. Amazon.com is an example of a recommendation system. This uses personalized data to make suggestions that a user may like. RS generates a list of recommendations by several methods: • Collaboration-based filtering method • Content-based method • Knowledge-based method • Hybrid-based method The hierarchical model of the recommendation system is given. below: Figure 1.1 Hierarchical model of the recommender system2. Recommendation Approaches Collaborative Filtering (CF) Based Approach System2.1Collaborative filtering based recommendation is a data filtering technique based on the collaboration of other users. Collaborative filtering uses the user-item matrix despite user or item information. Collaborative filtering is the most used and well-known recommendation technique, widely used due to its simplicity and good results. The first recommender system, Tapestry [5], used this term collaborative filtering, and since then it has been widely accepted. It is based on the fact that if two users X and Y rated n items the same or behave similarly in any environment, they will also rate or behave the same way on other items. Collaborative filtering is divided into two groups: • Memory-based: Memory-based b...... middle of paper ...... the sic system requires knowing the album, artist, singer, composer, etc. The content-based recommender system fails to give useful recommendations if the content does not include a sufficient amount of information to differentiate items the user likes from items the user does not like.2.3 System Hybrid-based recommenderHybrid-based recommender systems merge two or more recommendation approaches to achieve better performance with fewer limitations of each. Generally, the approach based on collaborative filtering is combined with another method in order to eliminate problems. Table 1.1 presents some of the hybrid methods used. Robin Burke (2002) proposes seven classes of hybrid methods: weighted, switching, mixed, feature combination, feature augmentation, cascade, and meta-level. Details are given in tabular form.