Behaviour Analysis using Statistical Learning Assisted Fuzzy Rough Set Theory (SL-FRST): A Comprehensive Investigation on Autism Spectrum Disorder (ASD)

Abstract

Autism Spectrum Disorder (ASD) indicates a developmental disorder where they face issues in communication, societal interaction, repetitive behaviour, and tampering. Every individual faces diverse weaknesses and strengths, and wide variations of severity or symptoms can be spotted. Behaviour analysis plays a prominent role in understanding the nature and treating the children. Examining behaviour can give a systematic approach for assessing, intervening, and understanding the behaviour issue related to the disorder. Incorporating statistical analysis and fuzzy-based behaviour analysis can assist ASD children. Overall, well-being and quality of life can be enhanced by analyzing behaviour. Initially, the data is scaled to normalize the features, and statistical analysis is applied. Multivariate Data Analysis (MDA) is utilized to identify the complex relationships and patterns from within the dataset. MDA permits exploring interrelationships among multiple variables, which interact with each other and can collectively impact the outcome. This research uses Principal Component Analysis (PCA) as an MDA technique. Analysis and predictions are more accurate and robust. Fuzzy Rough Set Theory (FRST) handles complex data, and the fuzzy sets are determined using the feature score. The fuzzy rules are generated to define the feature relation, which assists in behaviour analysis. The experimental results of the proposed Statistical Learning Assisted Fuzzy Rough Set Theory (SL-FRST) outperform the existing state-of-the-art technique.