Decision tree optimization for Sukuk rating prediction

Abstract

Sukuk, an Islamic finance framework for securitization, has increasingly popular in the last few years. The fast growth of Sukuk as Shariah-compliant financial instruments alternative to conventional bond has raised the issue of rating the Sukuk issuance. Recent trade and academic literature show that predicting Sukuk rating has been of interest to potential investors as well as to the firm. Unfortunately, in the current practice, often conventional bond and Sukuk were rated using similar method, hence ignore the fact that these two instruments are different in nature. It is the aim of this research to develop an optimum model to predict Sukuk rating using decision tree approach. Several models were produced using different attribute selection measures, namely gain ratio, Gini index and information gain. The effectiveness of the proposed models were evaluated using dataset on Sukuk issuance for domestic from 2006 to 2015. The results indicate that the decision tree model with Gini index as the criterion performs significantly better than the model produced using decision tree algorithm.

Keywords: Sukuk, rating, prediction, neural network