Predicting Autism in Children

Abstract

Our project aims at designing such a system which can predict the autism of a person given the reports of doctor with symptoms. In this project, three different models such as Logistic Regression, Support Vector Machine and Naive Bayes(through weka) have been investigated for the sole purpose of predicting autism. The data set consists of doctor reports whether the patient is autistic or non autistic . The features include the prediction of the autism given doctor reports and the symptoms from which they are suffering like fever,surgery,age,speech etc. In this project, we have studied several different ways of forming up input data sequences, as well as different architectures that may lead to effective prediction of autism. Using the proposed Machine Learning Model, we show that the f1 score and accuracy score of the logistic regression are 0.93 and 0.93. Our test outcomes have demonstrated that Machine Learning might be utilized for effectively foreseeing the autism in a person. For further expanding the execution of the anticipating calculations, earlier data about every patient would be alluring and the parameters of SVM and logistic regression could likewise be tuned