Developing Financial Distress Classification Approach based on Fuzzy-Rough Nearest -Neighbor approach

The traditional statistical models assume the covariance matrices for two populations are identical and both populations need to be described by multivariate normal distribution. Clearly, these assumptions do not always reflect the real world. Rough set is a new technique for data mining domain application. In this study, establish a financial distress Classification model using Fuzzy Nearest–Neighbor and Fuzzy-Rough Nearest–Neighbor approach to capture fuzzy rule form data sample. The empirical research which is based on the latest data of Taiwan’s listed company, the result shows that this method is high accurate and have reinforcement learning properties and mapping capabilities. The result shows that for this Taiwan listed company financial distress classification problem, fuzzy-rough NN approach offers a viable, if not the best alternate approach for our research design compared with fuzzy NN.