Hybrid Classifier for Predicting Financial Distress

The challenges and competition in the investment world recently became a great focus to be studied as it is greatly linked to profitability. Business sustainability has become a great issue to ensure the profit generated and keep the operation going, reducing the risk of financial distress of a firm. Model to predict the financially distressed firm has been developed over decades, using conventional statistical and recent data mining methods. A combination of clustering and classification methods called hybrid classifier, used in this paper. Cluster analysis, using k-means, conduct as the pre-classification. The clustering result used to construct the classification model into two clusters. From the cluster analysis, 17% of overall data clustered into cluster 1, meanwhile, 83% into cluster 2. The clustering results later used as the output label to the classification phase. The result shows a great overall accuracy of 98.6111% from the logistic regression and C4.5 decision tree completed with boosting, with an AUC value of 0.996 and 0.990 classified as excellent classification.