Design and Development of Machine Learning and Evolutionary Computation Methods for Risk Factors Identification in Early Childhood Disability

Abstract

Timely identification of social and emotional disorders is crucial for the immediate welfare and future well-being of young children. The present study encounters challenges in identifying and establishing risk factors associated with early childhood impairment. Consequently, the overall system performance is substantially reduced. In order to tackle the aforementioned challenges, this study proposes the utilisation of the Cuckoo Search Optimisation with Adaptive Network-based Fuzzy Inference System (CSO+ANFIS) technique. The aim is to effectively generate and identify risk factors associated with early childhood disability. This study employs the CSO method to select the most important attributes and determines the optimal objective function based on the highest fitness values. The ANFIS technique focuses on identifying important risk variables by analysing the hidden layer and fuzzy inference values. The experimental results have shown that the suggested CSO+ANFIS technique surpasses the current paradigm in terms of accuracy and sensitivity metrics.