Data Mining Approach to Prediction of Going Concern Using Classification and Regression Tree (CART)
This paper has employed a data mining approach for Going Concern Prediction (GCP) for one year ahead and has applied Classification and Regression Tree (CART) and Naïve Bayes Bayesian Network (NBBN) based on feature selection method in Iranian firms listed in Tehran Stock Exchange (TSE). For this purpose, at the first step, using the Stepwise Discriminant Analysis (SDA) has opted the final variables from among of 42 variables and in the next stage, has applied 10-fold cross-validation to figure out the optimal model. McNemar test signifies that there is a significant difference between the two models in terms of prediction accuracy and CART model is able to predict going concern more accurately. The CART model reached 99.92 and 98.62 percent accuracy rates so as to training and holdout data.
