Penerapan Deep Learning dalam Pendeteksian Autism Toddler

Diah Ayu Ambarsari, Ridan Nurfalah, Sandra Jamu Kuryanti

Abstract

Health is a very important thing. Everyone can overcome health problems. Children's health is the dream of every parent. During the growth period the child will switch several times which can stop their development. Parents must be more sensitive and have extensive knowledge in health. The problem that often occurs is that parents do not know the initial autism symptoms that occur in the baby, so more parents assume if it is okay, this situation accelerates the diagnosis process, whereas autism disorders can be detected early by looking at growing habits child development every time an autism transfer is a developmental development in children, autism must facilitate quickly, because with autism treatment quickly and quickly will help autistic patients grow back to normal. To help understand the children mengamalim autism, the authors conducted research with new methods. In a previous study, Fades Tahbatan conducted research to ascertain whether the child was autistic or not using a tool. But it only produces data sets., It turns out to have attributes that are not yet precise, which increases the level of accuracy. In this research, use the method of deep learning and improve accuracy, the application used is fast miners. The variables are then processed so as to produce a prediction model from the data set obtained. Accuracy values that can be processed are sufficient while accuracy = 98.96% precision = 96.74%, recall = 98.49% with AUC of = 0.90 Keywords: Autism, deep learning, toddlers  

Full Text:

PDF

Refbacks

  • There are currently no refbacks.