In the world of technology, tools and gadgets, a huge amount of data is produced every second in applications ranging from medical science, education, business, agriculture, economics, retail and telecom. Higher education institutes play an important role in the overall development of any nation. For the successful operation of these institutions, continuous monitoring for improving the quality of education and students is required, which is the subject of this article. A huge amount of data that education systems produce increases every year and it is difficult by traditional techniques to manage, predict and analyze this data. This challenge can be addressed through mining large amount of data. It enables the institutions to use their present reporting trends to unmask hidden patterns and identify data relationships. Through this, institutions easily predict which students are likely to dropout, and their performance. Present paper conducts a detailed and exhaustive study on techniques and approaches implemented in education mining for predicting dropouts.