Autism Screening on adults by developing a Custom ANN model using neumeric data
A custom ANN model developed to predict the likelihood of autism on adults using a dataset called ""Autism Screening on Adults"" which has 704 person, 21 datapoints including survey and demographic variables. It has explored Autism across age, gender , ethnicity, country of residence and other variables. 3 custom Artificial Neural Networks developed, whereas model-3 outperformed others.Architecturally, It consists of a input layer, followed by a dense layer which is fully connected with the input layer, a batch normalization layer to mitigate internal covariate shift and a dropout layer to prevent overfitting. After that, 2 dense layers with a dropout layer. Next, There is a dense layer, followed by a batch normalization layer and a dropout layer. Again, another dense layer, from there on a batch normalization layer and a dropout layer and a dense layer has been implemented.