WORD BASED LANGUAGE MODEL USING LONG SHORT TERM MEMORY FOR DISABILITIES
1Guruvammal.S, Dr.T.Chellatamilan, Dr.L.Jegatha Deborah
Identification of Autism Spectrum Disorder (ASD) among children plays a vital role to develop their social interaction through early intervention programme. Our proposed model categorizes the children who affected with ASD from typically developed children based on their linguistic knowledge. Existing methods are lagging with short-term memory and early saturation of learning rate. Our proposed model solves the demerits of existing approaches and has the advantages of classifying the children with ASD accurately with high precision and recall as well as a new model of LSTM based deep learning neural network has used. Word prediction is useful for disabled persons and IQ of Autism children will be tested based on latent semantic analysis. The experimental results showed in this paper promisingly.
-Autism Spectrum Disorder, Long Short-Term Memory, Language Modeling, Natural Language Processing