Analysis of Jordanian University Students Problems Using Data Mining System

Many social media users, especially university students, suffer from some problems that occur in their daily lives, such as anger, stress, depression, sleep problems, physical and psychological fatigue. Most of the students do not express their feelings to their parents and friends. Therefore, social media such as Facebook and Twitter are the best way to express their feelings and opinions because they feel safe and more comfortable there. This paper attempted to collect posts from Facebook and Twitter tweets. In this research, we focused on Twitter tweets and Facebook posts, which were mainly in the Arabic language and belonged to the Jordanian University Students. The huge amount of tweets and posts were not clean, so tweets and posts needed to be analyzed by labeling them to study load, sleep issues, negative feelings and positive feelings. Then, we classified them using many algorithms in data mining, such as a Naive Bayes Multi-Label Classification Algorithm (NB), K-Nearest Neighbor Classification Algorithm (K-NN) and Decision Tree Classification Algorithm. The results were compared based on three parameters which are: accuracy, precision and recall. The NB Algorithm achieved the best performance in terms of accuracy, precision and recall value, in comparison with the K-NN and Decision Tree Algorithm.