Illustration of the hospitalization (left) and ASD (right) datasets. For each ASD patient, we created a vector from the frequency of occurrence of each concept (C-1, C-2…) mentioned in their medical notes, ICD9 codes associated with a visit (ICD9-1, ICD9-2…) and medications prescribed (DRUG-1, DRUG-2…) within each 6 month period of their medical history captured in our database.