This thesis aims to study multi-dimensional association rules and data visualization for the schizophrenia outpatient data which are recorded in ICD-10 format. The data comprised of 4 main dimensions: psychoactive substance use, duration of substance use, patient demographic data, and schizophrenia status.
The total 24 datasets were extracted from those main dimensions which are specified to the schizophrenia as the target.
The processed datasets were discovered the associations by the Frequent Patterns-Growth algorithm (FP-Growth), which is the extension of the traditional one, the Apriori algorithm, to determine the rules representing the association of schizophrenia with the other factors. The result revealed the association rules between schizophrenia and substance use in 5 main features: 1) Thai men 35-44 years old, 2) Bangkok as residence, 3) nicotine use history, 4) amphetamine use history in a period of 7-12 months, and 5) diagnosis of substance dependence. All of the selected association rules had an acceptable confidence level over 0.90. This could confirm that the multi-dimensional association rules driven by the FP-Growth algorithm could be an appropriate technique to demonstrate a relationship pattern of data from the large database.