Current Proceedings on Technology
Yazarlar: Mohsen Shokoohi-Yekta, Maryam Mahmoudi, Bagher Ghobari- Bonab, Amir Ali Bagherzadeh, Hadi Moradi, Hamid Reza Pouretemad, Saeed Akbari Zardkhaneh, Salahadin Lotfi
Konular:-
Anahtar Kelimeler:Autistic disorder,Diagnosing autism,Expert system,Computerized screening
Özet: Prevalence of autism is growing, and it is important that they be diagnosed early. The aim of the present study is to provide an expert system software program to help children being diagnosed with autistic disorder. A questionnaire was developed and completed by 85 parents in the Center for the Treatment of Autistic Disorders, 65 parents of normal children and 20 parents of children with Down syndrome. After the completion of the long form of the questionnaire, a short form with more important items was formed through machine learning techniques such as Support Vector Machine and Random Forest. Finally, to study the validity and reliability of the expert system, machine-learning methods, Kappa agreement and Cronbach’s alpha were used. Findings indicate that the expert system has satisfactory levels of reliability and criterion validity. Therefore, it can be used for screening children with autism at the early ages of 2 to 6 years old.