An Automated System to Diagnose the Severity of Adult Depression

Depression is a common but ominous psychological disorder that threatens one\´s quality of life. The screening and grading of depression is still a manual process and grades are often determined in ranges, e.g., "mild to moderate\´ and "moderate to severe\´ instead of making them more specific as "mild\´, "moderate\´, and "severe\´. Such grading is confusing and affects the management plan. Given this practical issue, the present paper attempts to differentiate depression grades more accurately using a Back Propagation Neural Network (BPNN) classifier, built in MATLAB. The overall accuracy of the classifier is 100% for mild, 77% for moderate and 90% for severe grades with a good model fit (R=94%).