Usage of cluster algorithms in health studies: An application

Abstract

Objective: The purpose of this study is to introduce different clustering algorithms and show how and which cases should be correctly used. At the same time, different clustering algorithms results which can be applied on a real data set were compared. Material and Methods: Individuals grouped to low, medium and high risk clusters with using different clustering algorithms by examining risk factors of cardiovascular disease. EM, Denstiy based clustering, K-Medoid, Cascade K-Means and K-Means used for evaluating risk groups. Results: According to the evaluations, for two different data sets the kappa coefficients were statistically significant and its degree are intermediate. In terms of both data sets the most convenient and fastest algorithm is farthest clustering algorithm. The results obtained by Make Density Based and EM algorithms gave the most accurate desicions in terms of the distribution of the groups among Framingham risk groups cross tables. Conclusion: As a result, with taking into account the criterion of clinical information it is thought that the examination of clustering of risk factors of the disease, will be played an important role for introduction of accurate disease diagnosis. In addition we believe that when considering data distribution and characteristics of data sets clustering algorithms can be used as a diagnostic tool for the plannings and diagnosis of diseases in the field of health.