An Enhanced Bankruptcy Prediction Model Using Fuzzy Clustering Model and Random Forest Algorithm

Abstract

The rapid advancements made in Information Technologies (IT) have evolved in the prediction model for financial data. Research in the prediction of bankruptcy is inclining to owe to the growth of related associations with economic and social phenomena. Financial crises in recent scenarios have influenced the growth of financial institutions. Hence, the need for bankruptcy risk prediction at an earlier stage is of prime importance. Though several prediction algorithms were suggested, the predictive models' accuracy is still a challenging task. In this paper, a bankruptcy prediction model is developed by integrating the Fuzzy clustering model and Multi-objective random forest classifiers. The voluminous number of records of the financial dataset and polish bankruptcy dataset is collected from a public repository. It is pre-processed using a MapReduce technique, one of the Big Data approaches. Benefits given by big data approaches help to achieve better flexibility towards a variable declaration. The collected records are pre-processed and organized under efficient index construction. FCM is employed to cluster the data for analytic purposes. Finally, a multi-objective Random Forest classifier helps to develop a prediction model for bankruptcy. Experimental analysis is carried out with accuracy, precision, sensitivity, and specificity compared with existing, Genetic algorithms. Compared to the existing technique, the proposed technique has obtained 80% accuracy.

[Retracted] Latest Lessons from the Bankruptcy of State‐Owned Enterprises (SOEs) in China: An Interpretative Structural Model (ISM) Approach

“…e human resource was the foundation of an enterprise's development. DMTG's workforce is large, and employees lack Unprofessional top management F 13 Top management corruption F 14 High debt F 15 Outmoded operation management F 16 Lack of staff motivation 16 enthusiasm for their work limiting their potential for further development [32]. Understandably, DMTG in human resources is not ready to keep pace with the pace of reform and opening up.…”

“…ese methods have been widely used, but the logic of dealing with complex problems is often not strict enough; therefore, some scholars choose quantitative analysis. Altman's Z-scores [10] are expounded by Ni et al [11]; these classical bankruptcy prediction models (for instance, Springate and Meachin [12], Saladi and Yarlagadda [13], and Legault and Kirton [14]) are widely used by researchers [15]. Meanwhile, structural models, e.g., the Merton, Black-Cox, Leland-Toft, Longstaff-Schwartz, flat barrier, and Geske models, are hot applications in bankruptcy prediction.…”

Penerapan Metode Oversampling SMOTE Pada Algoritma Random Forest Untuk Prediksi Kebangkrutan Perusahaan

“…Kebangkrutan merupakan penutupan perusahaan atau disebut juga likuidasi, hal ini merupakan kondisi terjadinya kegagalan dalam masalah keuangan perusahaan yang tidak sebanding dengan jumlah asetnya. Kerugian tersebut dapat berdampak pada pemegang saham, karyawan, pelanggan dan investor [2]. Oleh karena itu prediksi kebangkrutan berperan penting dalam bidang ekonomi dan keuangan.…”

“…Proses prediksi adalah membuat keputusan yang berhubungan dengan operasi bisnis dengan cara yang dapat diandalkan [2]. Tujuan dari prediksi kebangkrutan adalah untuk menciptakan model prediksi yang efektif untuk mengidentifikasi risiko kebangkrutan berdasarkan data keuangan yang dimiliki oleh perusahaan.…”