Autism is a category of beginning sign of deficiency in early phase of life and difficult to detecting, particularly in small children. This is for the reason that the autism symptoms are depend on the reaction of children to their cognitive function. On the other hand, if autism is not distinguished and delighted among the age of 2–5 years. The treatment next to a later on next phase, it becomes more complicated. Furthermore, a lot of parents are not capable to articulate each and every one of the symptoms, that children is familiarity. In this article, planned a method for indicating the probable symptom starting the beginning indication through with the preceding extended expression patient report to identify the appropriate autism category. Here the proposed method applies equally the used machine learning algorithm as well as confabulation speculation to get ready a medium which assist for receiving each category of relations for producing the rules. The clarity is used to calculate the frequencies of occurrence on a couple of provisional objects. Hence it is used for coupling of symptoms to forecast the subsequently the most feasible and convinced symptom for appropriate identification of autism. By using this method, it can identify both the ordinary and uncommon type of autism. For observing the autism identification contrivance uses a lesser amount of reminiscence for its implementation. It is quicker for the reason that single instance database admittance as compare by using the Apriority algorithm.