Public Nor Bazilah Safiar
Sabri Ahmad
Jusoh Yacob
Mustafa Man
Universiti Malaysia Terengganu
Currently, The use of statistical models to predict business failures have been noticed.However, very few studies have been conducted to predict the loss of a co-operative business. The objective of this study to construct the Logistic Regression model(LR) and Neural Network (NN), to compare and to identify the best model (LR and NN) and identify important variables in forecasting losses cooperative. Data from 2009 until 2011 were obtained from the Department of Cooperatives Commission of Malaysia (CCM) Terengganu branch were analyzed using SPSS Clementine version 12.0 Results shown that the Prune method in the NN model gave the highest value of testing 82.77%, the sensitivity of 82.90%, the specificity of 75% and minimum error (MSE) of 17%. Capital was the most variable influence towards the cooperative loss due to the highest value of 0.23(23%).
