Experimental Framework: Bankruptcy Prediction Using Soft Computing Based Deep Learning Technique

Abstract

The experimental framework consists of predicting bankruptcy through soft computing based deep learning technique. Deep learning has evolved a promising machine learning technique in past few years [82]. It is deep structured learning and concerned with ANN study containing more than one hidden layer. It is based on composition of several layers with nonlinear units toward feature extraction and corresponding transformation. Here each preceding layer provides input to successive layer. Deep learning algorithms are executed in either supervised or unsupervised manner. These algorithms learn from multiple representation levels corresponding to several abstraction levels. Deep learning has been successfully applied toward several categories of pattern recognition and computer vision problems with considerable success.