Prediction of autism severity level in Bangladesh using fuzzy logic: FIS and ANFIS

Abstract

A type of neurodevelopment disorder also known as autism is currently more visible than before among the people of Bangladesh. Some research works could be found on autism but very few papers are guided to measure the severity level. Hence, this research focuses on attaining the severity level of autism using fuzzy methods like Mamdani Fuzzy Inference System (MAMFIS) and Adaptive Neuro-Fuzzy Inference System (ANFIS). A survey has been conducted on autistic children to find the severity level. The levels used in this research are low, medium, high. A comparative study of those two methods has been reported in this paper. By using ANFIS we get better accuracy compared to the FIS model.