Clustering CST Telephone Bill Using Hybrid Pcak Algorithm

Abstract

In this paper, we chose to study Principal Component Analysis (PCA) and K-means clustering algorithm (K) to investigate a set of real-world telephone data. The raw data we received showed high variation between maximum and minimum data. A hybrid PCAK approach was thus proposed. The PCA normalizes the data range and reduces its dimensionality. While K-means clusters the normalized and dimension-reduced data into k clusters. The clustered output from PCAK showed the telephone usage patterns of CST staff.