Exponential Method for Determining Optimum Number of Clusters in Harmonic Monitoring Data
Clustering is an important process for finding and describing a variety of patterns and anomalies in multivariate data through various machine learning techniques and statistical methods. Determination of the optimum number of clusters in data is the main difficulty when applying clustering algorithms. In this paper, an exponential method has been proposed to determine the optimum number of clusters in power quality monitoring data using an algorithm based on the Minimum Message Length (MML) technique. The optimum number of clusters has been verified by the formation of super-groups using Multidimensional Scaling (MDS) and link analysis with power quality data from an actual harmonic monitoring system in a distribution system in Australia. The results of the obtained super-group abstractions confirm the effectiveness of the proposed method in finding the optimum number of clusters in harmonic monitoring data.