Construction Financial Crisis Warning Model Using Data Mining
This paper employs artificial neural network of data mining and decision tree algorithm to build financial crisis warning model. The research results show that, forecasting performance of artificial neural network is better than that of decision tree model, hence, “financial statement average warning model” established through artificial neural network based on the average revenue of the past three years before financial crisis has better forecasting performance than the “annual report forecast model”. Factor analysis is employed to select common factor in 1 year before financial crisis, and the critical variables of financial crisis are found to be: debt-to-equity ratio, quick ratio, borrowing dependence, inventory turnover ratio, and earnings per share. According to the decision tree rule, variables differentiable to financial crisis warning are debt-to-equity ratio, earnings per share, and borrowing dependence.
N. H. Pan et al., "Construction Financial Crisis Warning Model Using Data Mining", Advanced Materials Research, Vols. 271-273, pp. 684-688, 2011