A Comprehensive Analysis of Traditional Clustering Algorithms On Corporate Bond Data

This paper examines various clustering algorithms for corporate bond. The corporate bond is a bond issued by the corporations with certain time limit after which the bond will get mature. In order to raise funding such as explanation of business and for many other reasons corporation issues bond to the organization. A few years ago, the corporate bond market had to face very poor clustered default events in comparison to those experienced throughout the recession. The purpose of bond clustering, any organization can choose the bond with high ratings and high yields. Clustering is the technique to group data by identifying the common characteristics on a large dataset. In the proposed work K-Means, Hierarchical, Self Organizing Map, Fuzzy-C Mean and Gaussian Mixture Model clustering algorithms have been used to segment the bonds. The Portuguese retail bank with data collected from 2008 to 2013 is used for the presented work. The proposed work results, K-Means gives the most efficient result in terms of time and accuracy.