Saving time is one of the most important things. A Recommender system certainly saves our time by generating a prediction of things for us from an extensive database that we might like. Due to the widespread use of the internet, it became a popular area of research in the machine learning community. It is being used in almost all web sites these days to keep track, what users want or like so that when they visit next time, it can predict something for them. A lot of research has been done on this field, but still, the performance of these systems is not that good to be accepted by the end-users. Among these, model-based collaborative filtering is widely used in commercial recommender systems due to its scalability and capability of handling sparse datasets. In this paper, we will provide a detailed insight into these techniques along with their advantages as well as disadvantages.