Performance Analysis of Student Healthcare Dataset using Classification Algorithm

Abstract

Nowadays health is considered as a backbone in terms of performance based on Internet of things (IoT devices), which turned out to be important in diagnosing health level of person with the type of disease a person is suffering with plus its severity level. Basically, IoT sensors operate on medical devices produce large volume of dynamic data. The fluctuation in health data, which forced to use data mining tools and techniques for extracting useful data. Therefore, for applying data mining techniques, heterogeneous data needs to be preprocessed. Therefore, by refining the collection of data, health parametric data mining yields better results with associated benefits. The decision tree is employed in order to consolidate the health attributes of the students to decide the level of health scale. This could leads to evaluate the level of performance of the student in class. After mining the student’s health data it is passed to K-Fold cross validation check, so that to determine the accuracy, error rate, precision and recall. This is considered as enhanced diagnosis method as fixed patterns for decision tree makes precise decisions. Considering a case study of students health prediction based on certain attributes and the level of attributes (Normality to study) diagnostic as pattern based using K-NN and decision tree algorithm, which are tested on trained dataset using WEKA tool. At the end, the comparison of different algorithms will be reflected to generalize the introduction of optimized classification algorithm.