Change Point Estimation Based on a Weighted Consensus Clustering Approach with Multiple Steps (A Real Case in Health Care: Diabetic Patients)

One of the primary objectives of control charts is to accurately detect the occurrence of changes in statistical processes. This enables process analysts to identify the factors responsible for the change and take corrective actions. The clustering method is a practical approach that has been developed to estimate the time of change. In recent years, a new approach called weighted consensus clustering (WCC) has been introduced in the field of clustering analysis. This approach can be utilized to cluster different trends in process parameters, resulting in a final clustering that exhibits higher quality and stability compared to the initial clustering. One of the emerging aspects in the health care field is the use of new statistical approaches to rapid disease diagnosis. In this paper, a novel change point estimation scheme called WCC has been developed to detect changes in the diabetic patients’ processes. In addition, to demonstrate the advantages of the proposed scheme, its performance is compared with an existing method known as maximum likelihood of clusters (CMLE) through simulation studies as well as a real case study in the healthcare system.