Clustering Diabetics Data Using M-CFICA

E-Health has grown popular due to a wide range of services provided. The role of a patient has also changed in today’s health care as they are expected to use ICT services to gain information and knowledge to know about their well-being. In the field of data mining clustering is a widely used technique for discovering patterns in underlying data. Traditional clustering algorithms are normally limited to handling datasets that contain either numeric or categorical attributes. However, datasets with mixed types of attributes are also common in real life data mining applications. In this paper a cluster feature based incremental clustering algorithm, MCIFA (Cluster Feature-Based Incremental Clustering Approach to mixed data) is applied on the diabetes dataset to check its suitability in the medical domain. The achieved clustering accuracy in results section shows that this is indeed suitable for medical domain and can be used for ‘e-prescribing’. But it needs to be fine-tuned so as to increase the clustering accuracy as the percentage of allowed error-rate in medical domain should be as small as possible.