Analysis of Bankruptcy using Data Mining Approach

This study involves the development of neural network prediction model to predict the stage of bankruptcy of a company. A total of 367 data was attained from the Registrar of Business and Companies, Kuala Lumpur Stock Exchange (KLSE) and Bank Negara Malaysia (Central Bank of Malaysia). The data was then analyzed by considering the basic statistics, frequency and cross tabulation in order to get more information about the data. Initially, the data was classified using logistic regression.In addition, it was also trained using neural network in order to obtain the bankruptcy model. The findings show that the most suitable prediction model consist of 12 nodes of input , hidden layer 6 node and one output layer. The generalization performance of the selected model is100%. This methodology should be able to provide some new insight into the type of pattern that exists in the data. Thus, neural network has a great potential in supporting for predicting bankruptcy