An Innovative Framework for Classification of ADHD using Machine Learning Algorithm

Precise analysis and diagnosis of Attention Deficit Hyperactivity Disorder (ADHD) is a major chore in medical field. Underdiagnosis leads to notable side effects in medical field. ADHD has critical features of disorder, hence still there is no exact estimatimated intelligent technique for the problem. Nowadays an analysis and diagnosis of disorder by Machine Learning (ML) technique has received an enormous recognition. The proposed method is a robust ML based framework which analyses attention and hyperactive/impulsive parameters as a crucial biomarker for the automatic diagnosis of ADHD. The proposed technique initially uses mining process to select 49 features (out of 225) as the significant predictors. These features include variety of age groups between 8 to 16. The framework integrates various subtypes testset phenotypic feature engineering observations with futuristic binary ML classification algorithm and crucial features were chosen and combined. The algorithm classifies each feature using decision tree ML algorithm. Important characteristics were chosen and then combined. The classification method was recursively repeated until a significant pattern for the model is found. Machine learning random forest classifier achieved F1-Score of 87.7% of an average accuracy on a classified dataset of 49 patients. The result outcome indicates that the union of complementary characteristics can lead to significant improvement in the performance of the predictive framework model. The innovative framework designed were able to provide the foremost classification using generalized decision tree. This work is designed using ML expert system and that reduces underdiagnosis of disorder. The framework is beneficial for the diagnosis and treatment efficacy. This study is the primary work in power BI using machine learning based method for the classification of ADHD patients.