Prediction of Autism Spectrum Disorder Using Rough Set Theory

Autism Spectrum Disorder (ASD) is a neurological disease that starts early in childhood and persists throughout a person's life. It is a condition linked with brain development and influences a person's behaviour and their interaction with others. Autism has a wide range of symptoms which can vary from person to person. There is no direct medical test to diagnose ASD disorder and hence trained physicians are needed to oversee the person's behaviour development to detect it. There is no cure for ASD, and early detection of the illness will be able to make significant quality improvements in the behaviour of the affected person. Machine Learning techniques are widely used to identify the factors associated with the disease, thus helping in early detection. This paper attempts to explore the possibilities of analyzing the autism data sets of adults using rough set theory and predict the main factors associated with the disorder for providing an early treatment. A comparative performance analysis of the results is done using two rough set algorithms, and the results indicate that the genetic algorithm gives a better performance in this domain.