Patient Data Clustering using Fuzzy C-Means (FCM) and Agglomerative Hierarchical Clustering (AHC)

Abstract

Generally, the current system development only include the input, view, and reports. At Jogja Hospital, a system with a patient database can only provide information about the percentage of male and female patients. Its unable to extract more specific information, even though medical record data has a lot of information. The complete information should be used as a reference for the authorities to make a decision. This information can be obtained by analyzing and processing the medical record data. One way to extract information from this data is clustering. The domain of this study is patient data. Before the data is clustered, preprocessing is needed through name standardization, numeration, and data normalization. During the clustering process, the algorithms used are Fuzzy C-Means (FCM) and Agglomerative Hierarchical Clustering (AHC). Two algorithms are implemented to determine which algorithm is the most appropriate and fast to handle the processing of patient data. The results of the study show that the processing time required to do clustering with FCM algorithm is relatively faster than AHC algorithm. For data with small volumes, the iteration of FCM algorithm is more than AHC algorithm, however, the results of the clustering using FCM algorithm are easier to interpret than AHC algorithm. From the visualization of clustering results, found that the cluster pattern with FCM algorithm is better based on the three variables used as references. So the most suitable algorithm to use is Fuzzy C-Means (FCM) for processing patient data.