enhancing efficiency of neural network model in prediction of firms financial crisis using input space dimension reduction techniques

the main focus in this study is on data pre-processing, reduction in number of inputs or input space size reduction the purpose of which is the justified generalization of data set in smaller dimensions without losing the most significant data. in case the input space is large, the most important input variables can be identified from which insignificant variables are eliminated, or a variable can be used in combination with several variables. this approach leads to reduction in number of inputs and input variances and improvement in results. this research intends to build the best neural network model using financial variables (the financial ratios profit and loss statement and balance sheet) and such techniques as mean equality test or independent samples test (ist) , multiple discriminant analysis (mda), and principal constituents analysis (pta) in order to reduce the input size and space and to enhance financial crisis model’s prediction power and eventually to aid better decision making on the part of users of financial statements in prediction of financial crisis. in this research, four financial crisis prediction models (artificial neural network (ann), combination of principal constituents analysis and artificial neural network model, combination of ist and ann, and combination of mda and ann) are used for prediction of financial crisis two years prior to its occurrence. next, given the obtained results, the models are compared with each other and the best model is extracted. considering the test results, use of the ist in construction of neural network model was found more efficient in prediction of firms’ financial crisis relative to other techniques investigated in this research.