Abstract: Social media has gained a lot of popularity amongst the students. Students share their feelings and their day to day experiences on it in a very informal and casual manner. Experiences and problems of students revealed through social media need human interaction or human analysis. Knowledge from such uninstrumented environments can present valuable data to report student problem. But mining knowledge from such data can be a very challenging task. The huge amount of data requires automated data analysis techniques. In this paper, a work-flow is developed which combines both qualitative investigation and large-scale data mining scheme. The data posted by students are collected and analyzed. It is found that certain issues like heavy study load, hectic schedule and lack of sleep are encountered by the students. Hence these issues are classified using Naive Bayes Multi-label Classifier algorithm. Here we are also using the clustering algorithm i.e. K-means. This both techniques classification and clustering can help in understanding the student’s problem in a very efficient way.

Keywords: Data mining, social media, text mining, social network analysis.