Bond Ratings Classification Models

Student Name: V.Fani Vamsidhar B.

This was the first study undertaken in my supervision for modeling CRISIL bond ratings. The methodology of attacking this problem was first streamlined through this study. In this project, Multiple Discriminant Analysis and Artificial Neural Network models were utilized for predicting CRISIL bond rating based on financial fundamentals of all the bonds rated by CRISIL for the calendar year 1996. Artificial Neural Networks were found to outperform the Multiple Discriminant Analysis models.