基于神经网络模型的上市公司财务预警研究

为了对上市公司财务状况进行合理预测,利用神经网络模型进行企业财务预警研究。首先,运用主成分分析等统计方法对初选的财务指标进行筛选,从初始指标中提取了五个因子作为财务预警指标;然后,选取2010至2013年间首次被特殊处理的上市公司作为财务危机样本企业,构建的财务预警模型提前三年的预测准确度近似于90%。研究结果表明,神经网络模型对我国上市公司财务状况预测具有较高的准确度,并且对预警指标进行适当的预处理,会提高模型预测的准确率。

Financial Crisis Warning Research Based on Neural Network Model

In order to increase forecasting accuracy, the paper uses neural network to establish a financial crisis warning model. Firstly, the paper filters 5 financial indexes with main-composition analysis and other statistical methods. Then, the paper chooses listed companies which had been special treated for the first time during 2010 to 2013 as financial crisis sample. After training neural network model by training data, the result indicates that the prediction accuracy of the neural network model approaches 90%. The research indicates that neural network is a relatively satisfying method for forecasting, and the model is more accurate by proper data pre-treatment.