Diagnosis and evaluation of ADHD using Naïve Bayes and J48 classifiers

Naive Bayes and J48 Data mining classifiers to Diagnose and Evaluate the Attention Deficit Hyperactivity Disorder is proposed in this paper. A sample questionnaire focusing the Behavioral, Educational and Medical characteristics of the children is prepared with three parts grouping the symptoms of different types of ADHD: Inattentive, Hyperactivity and Impulsive. School going children at the age of 5-9 years is targeted and an ADHD data set with 105 data samples with 30 attributes is created from their parents and teachers under the guidance of physicians. The ADHD data set is preprocessed to remove duplicate records, missing data, noisy and inconsistent data and classified initially into NOADHD and ADHD. The ADHD prone samples are further classified into moderate ADHD (ADHDmod) and high ADHD (ADHDhigh) classes using Naive Bayes and J48 classifiers and their classification performance is evaluated with the aid of Confusion matrix, ROC curve and Accuracy measures using WEKA tool. The experimental results show that the J48 classifier gives 100% accuracy on ADHD dataset compared to Naive Bayes which yields 94.3% with less time. To create awareness about the influence of ADHD among the school going children is the main intension of this work.