An efficient system to deduce depression using case-based reasoning and CHARM mining algorithm

This paper introduces a system for psychologists in assessing cases of depression and determines level of severity and finds all frequent symptoms using CHARM mining algorithm. This system has a user interface that requires the client's to answer multiple questions that corresponds to a specific symptom. The individual ratings and overall ratings of the symptoms go to a database containing data and diagnosis of patients for comparison. The existing system was implemented using rule-based reasoning. Due to some drawbacks case-based reasoning is used for comparing new cases with existing cases. Value difference metric function is used to retrieve nearest neighbor of the new case and the diagnosis of the nearest neighbor. With the founded symptoms, a large database is designed in order to find all frequent symptoms and it is used to reduce unnecessary symptoms from the database.