A machine learning approach to predict university students Hookah Smoking (HS)

Abstract

In recent years, Hookah (Shisha) has spread in general and large, including classical and electronic types, which have spread especially among young university students. the dataset was used on the university student at the University of Kirkuk, which was collected and analyzed using a special questionnaire about Hookah Smoking (HS) in the university students. The aim of this work is to find out how much the students are concerned about the recent Hookah Smoking in the university students and the extent of their consumption of the time that the student is supposed to devote to his studies at the college. Using the algorithms and techniques of data mining and machine learning to Hookah Smoking (HS), used decision tree and random forest algorithms to classify hookah smoking for university students. Then predict when the students smoke shisha and the negative impact of this time on the health of the university student, which in turn negatively affects his scientific level. Furthermore, best algorithm archive random forest has high classification rate than decision tree. New predictions can also be made for the development of statistics and tables that determined the type and quantity of consumption of Hookah Smoking and other side effects.