摘要
在司法实践中,众多领域都已经涉及精神损害抚慰金,受到学者们的广泛关注.现阶段的研究大多处于定性分析,定量分析尚处在起步阶段.该文以中国裁判文书网上的交通事故民事纠纷案为例,利用分层随机抽样方法获取1539份判决书作为样本,提取了13个影响精神损害抚慰金的变量,构建了逐步回归模型和随机森林回归模型,对精神损害抚慰金的特征进行了分析,并对模型的预测效果进行了评价.该文通过实证发现:①伤残程度、受害方精神损害诉讼数额、受害方物质损害诉讼数额、受害方物质损害获赔数额是影响精神损害抚慰金的主要变量;②伤残程度很大程度上决定了精神损害抚慰金,二者具有较高的正相关关系;③基层、中级人民法院中随机森林模型的决定系数(R^(2))值(0.671、0.699)和逐步回归模型的R^(2)值(0.716、0.770)相差不大;但是基层、中级人民法院中随机森林模型的平均绝对误差(E_(MA))值(0.405、0.390)比逐步回归模型的E_(MA)值(0.997、0.984)要小很多,总的来说,随机森林回归模型的预测效果比逐步回归模型预测效果好.该文研究希望有助于精神损害赔偿制度的理论探讨和法治体系的完善.
In judicial practice,compensation for mental damage has been involved in many fields,which has been widely concerned by scholars.At this stage,most of the studies are in the qualitative analysis,while the quantitative analysis is still in the initial stage.This paper takes the traffic accident civil disputes on the Internet of Chinese judgment documents as an example,and stratified random sampling method was used to obtain 1539 judgments as samples,13 variables affecting mental damage compensation were extracted,a stepwise regression model and a random forest regression model were constructed,the characteristics of mental damage compensation were analyzed,and the prediction effect of the model was evaluated.The empirical findings of this paper are as follows:①the degree of disability,the amount of the injured party's mental damage lawsuit,the amount of the injured party's material damage lawsuit,and the amount of the injured party's material damage compensation are the main variables affecting the compensation for mental damage;②the degree of disability determines the compensation for mental damage to a large extent,and the two have a high positive correlation;③The R^(2)values of the random forest model(0.671,0.699)and the stepwise regression model(0.716,0.770)in the primary and intermediate people’s courts were not significantly different;however,the mean absolute error(E_(MA))values(0.405,0.390)of the random forest model in the primary and intermediate people’s courts are much smaller than those of the stepwise regression model(0.997,0.984).In general,the prediction effect of the random forest regression model is better than that of the stepwise regression model.Hope to contribute to the theoretical discussion of the mental damage compensation system and the improvement of the rule of law system.
作者
年银果
潘颖群
王文强
NIAN Yinguo;PAN Yingqun;WANG Wenqiang(School of Mathematics and Computational Sciences,Xiangtan University,Xiangtan 411105,China)
出处
《湘潭大学学报（自然科学版）》
CAS
2024年第5期59-68,共10页
Journal of Xiangtan University(Natural Science Edition)
基金
国家自然科学基金(12471391)。
关键词
精神损害抚慰金
交通事故
民事案件
定量分析
预测模型
compensation for mental damage
traffic accidents
civil dispute cases
quantitative analysis
prediction models
