论文部分内容阅读
企业财务困境预测是财务实务界和理论界关注的热点问题之一。为了有效识别出在未来两年内有可能陷入财务困境的企业,该文提出了一种遗传算法优化灰色案例推理的新方法进行企业财务困境预测,并采用实证研究予以验证。理论研究中主要构建了财务困境预测的企业案例描述、基于灰色相似度的k近邻案例检索、基于相似度加权投票组合的评价结果集成和基于遗传算法的案例特征权重向量优化等四个关键内容。实证研究中,收集了270个上市公司ST前一年和前两年的数据为初始样本,通过格点搜索技术进行参数优化,采用余一交叉验证准确率作为评价标准,通过与多元判别分析、Logistic回归、BP神经网络、支持向量机的预测结果比较发现:该方法在企业财务困境预测中的准确率有较大提高。
Prediction of corporate financial distress is one of the hot issues that financial professionals and theorists pay close attention to. In order to effectively identify the enterprises that may be in financial distress in the next two years, this paper proposes a new method of genetic algorithm to optimize gray case inference to predict the financial distress of enterprises and verify it with empirical research. In the theoretical research, four key contents, namely business case description of financial distress prediction, k nearest neighbor case retrieval based on gray similarity, integration of evaluation results based on similarity weighted voting portfolio and case weight vector optimization based on genetic algorithm, are mainly constructed. In the empirical study, the data of 270 listed companies in the previous year and the previous two years were collected as the initial samples, the parameters were optimized by the lattice search technique, the cross-validation accuracy of Yuyi was used as the evaluation criteria, and by multiple discriminant analysis, Logistic regression, BP neural network, support vector machine prediction results show that: the accuracy of this method in the prediction of corporate financial distress has greatly improved.