A comparison of effectiveness of risk data clustering method in Psychiatric Patient Service

Abstract

In this paper, we clustered clinical risk data of a mental health service, Khon Kaen Rajanagarindra Psychiatric Hospital. This study aims to compare performance values of cluster (k) in k-means clustering algorithm and hierarchical clustering algorithm. The result shows that for k-means clustering algorithm, sum of squared error (SSE) is 32.68, minimum of distance (MD) is 1.38, mean squared error (MSE) is 2.95 and values of k is 11. Therefore, we found that k-means clustering algorithm is the most appropriate method for using in cluster the risk group of the Psychiatric Patient Service. The result also suggests that the most risky age is between the ages of 32 and 36. The result can be a guideline for further research about data prediction. The implications of this study can assist medical staff to be knowledgeable about what should beware of when they treat psychiatric patients and this can be basic planning medicate guidelines for medical staff.