| 978 | 6 | 496 |
| 下载次数 | 被引频次 | 阅读次数 |
财务预警是预测企业财务风险的重要手段。一方面，财务预警领域产生了丰富的研究成果，其潜在信息值得在统计模型中发掘应用。另一方面，由于不同行业间的财务结构存在差异，财务指标数据可能呈现异质性特点。为利用财务预警领域既有丰富成果来构建先验信息，并同时考虑数据的异质性，本文构建融合先验信息的整合财务预警模型。模拟实验证明，融合先验信息的整合模型在变量筛选和预测上具有明显优势。进一步的实证研究中，通过对财务预警领域核心文献进行文本挖掘，基于词频统计构建财务指标的先验信息集，并在食品饮料制造业、金属制造业、批发零售业三个行业数据上建立整合预警模型。结果显示，本文构造的融合先验信息的整合财务预警模型可以筛选出关键预警指标，具有较高水平的预测精准度，体现出本文模型的实际应用价值。
Abstract：Financial risk warning is an important approach to predicting the financial risks of enterprises. On one hand, the plenty of researches on financial risk warning, can be exploited further in statistical models. On the other hand, different industries have different financial structures, which leads to the heterogeneity of financial indicator data. In order to utilize the rich past literature and also consider the data heterogeneity, we propose an integrative financial risk warning model with prior information.Demonstrated by a variety of simulations, the proposed model shows good performance in both variable selection and prediction. In real applications, we first construct a prior information set based on frequencies of covariates in previous literature. Then the integrative model is built with data of three industries(food and beverage manufacturing, metal manufacturing, and wholesale and retail). Results show that, the proposed model can select the effective covariates for financial-risk warning and also achieve high prediction accuracy.
[1]方匡南,章贵军,张惠颖.基于Lasso-logistic模型的个人信用风险预警方法[J].数量经济技术经济研究, 2014(2):125–136.
[2]蒋昕艳.论加权平均净资产收益率的缺陷及改进措施——基于上市公司财务报表分析[J].吉林省经济管理干部学院学报, 2015, 29(4):40–42.
[3]李扬,李竟翔,马双鸽.不平衡数据的企业财务预警模型研究[J].数理统计与管理, 2016(5):893–906.
[4]马双鸽,王小燕,方匡南.大数据的整合分析方法[J].统计研究, 2015, 32(11):3–11.
[5]马永义.如何透过流动资产周转率研判企业基本面[J].商业会计, 2021(22):24–26.
[6]斯介生,李扬,谢邦昌.基于异质性数据的Logit变量选择模型研究[J].统计研究, 2017, 34(12):110–118.
[7]王翰雄,杨东华.资产负债率公式的理解与运用[J].财会月刊（理论版）, 2014(3):102–104.
[8]王乐,韩萌,李小娟,等.不平衡数据集分类方法综述[J].计算机工程与应用, 2021, 57(22):42–52.
[9]王小燕,袁欣.基于惩罚组变量选择的COX财务危机预警模型[J].系统工程, 2018, 36(3):113–121.
[10]王小燕,张中艳.带网络结构的自适应Lasso财务风险预警模型[J].数理统计与管理, 2021, 40(5):888–900.
[11]吴星泽.财务危机预警研究:存在问题与框架重构[J].会计研究, 2011(2):59–65.
[12]尹夏楠,鲍新中.企业财务风险测度与预警系统软件的设计与应用——基于行业差异视角[J].财会通讯, 2019(2):109–112.
[13] Chawla V N, Bowyer W K, Hall O L, et al. SMOTE:Synthetic Minority Over-sampling Technique[J]. The Journal of Artificial Intelligence Research, 2002, 16(1):321–357.
[14] De Smedt T, Daelemans W. Pattern for Python[J]. Journal of Machine Learning Research, 2012, 13(1):2063–2067.
[15] Jiang Y, He Y, Zhang H. Variable Selection With Prior Information for Generalized Linear Models via the Prior LASSO Method[J]. Journal of the American Statistical Association, 2016, 111(513):355–376.
[16] Li Y, Li R, Qin Y C, et al. Integrative Interaction Analysis Using Threshold Gradient Directed Regularization[J]. Applied Stochastic Models in Business and Industry, 2019, 35(2):354–375.
[17] Li Y, Wang F, Wu M, et al. Integrative Functional Linear Model for Genome-wide Association Studies with Multiple Traits[J]. Biostatistics, 2022,23(2):574–590.
[18] Liang W, Ma S, Zhang Q, et al. Integrative Sparse Partial Least Squares[J]. Statistics in Medicine, 2021, 40(9):2239–2256.
[19] Lin C C, Ng S. Estimation of Panel Data Models with Parameter Heterogeneity when Group Membership is Unknown[J]. Journal of Econometric Methods, 2014, 1(1):42–55.
[20] Ma S, Zhang Y, Huang J, et al. Integrative Analysis of Cancer Prognosis Data With Multiple Subtypes Using Regularized Gradient Descent[J].Genetic Epidemiology, 2012, 36(8):829–838.
(1)因篇幅所限，不同样本量设定下不同方法的对比结果以附表1展示，见《统计研究》网站所列附件。下同。
(1)CSMAR网址为https://www.gtarsc.com/。
(1)因篇幅所限，解释变量相关性情况以附图1展示。
(2)因篇幅所限，各方法变量选择情况对比以附表2展示。
基本信息:
DOI：10.19343/j.cnki.11-1302/c.2024.05.012
中图分类号:F275;TP277
引用信息:
[1]王菲菲,贾珂,张开宇,等.融合先验信息的整合财务预警模型研究[J].统计研究,2024,41(05):137-149.DOI:10.19343/j.cnki.11-1302/c.2024.05.012.
基金信息:
国家自然科学基金青年项目“大规模非随机分布数据下广义线性回归模型的理论与应用研究”（72001205）;国家自然科学基金面上项目“变量选择不确定性评价方法及其在管理科学中的应用”（72271237）; 教育部人文社会科学重点研究基地重大项目“数字经济驱动高质量发展的统计测度与分析研究”（22JJD910002）
2024-05-25
2024-05-25