<![CDATA[Bankruptcy is a condition while a company fails either economic failure or even financial failure. Bankruptcy causes a general seizure of all the assets of a bankrupt Debitor (company) that settled and managed by the Curator (supervisor of Debitor's asset). Because it can causes a severe consequences, several attemps were done as an alternative for bankruptcy prevention. One of those attemps is by predicting the bankruptcy itself. Backpropagation is a method of artificial neural network that widely used in the context of classification or regression datasets, one of the regression problem is prediction, because backpropagation is one of the supervised learning algorithm which the output or input values already known. In this study, backpropagation works for predicting the bankruptcy with Altman's five variabels as inputs and the results of Z-Score calculation as output target. The entire test that has been done produces the best MAPE value with average at 0,062% using learning rate parameter value at 0,2, 1000 iterations and 6 neurons in the hidden layer. This MAPE value is under 10% and close to 0% which included in the criteria of prediction with very good accuracy.]]>
Copyrights © 2019
