基于机器学习的企业价值影响因素分析

Analysis on Influencing Factors of En-Terprise Value Based on Machine Learning

本文选取的数据为2009年~2021年创业板企业年度数据,分为八大类因素搭建了创业板企业价值评估指标体系。综合使用SVR、GBRT、BP神经网络多个机器学习方法,先验证了机器学习对创业板企业价值预测效果,然后选择效果最优的机器学习模型对特征重要性进行排序,最后从机器学习角度综合分析了创业板企业价值的影响因素。结果表明:企业规模、企业盈利能力、股本扩张能力、发展能力以及创新能力类指标对创业板企业价值有较大影响,其中代表企业规模的资产总计指标对创业板企业价值影响最大。

The data selected in this paper is the annual data of GEM (Growth Enterprise Market) enter-prises from 2009 to 2021, which is divided into eight categories of factors to build the GEM en-terprise value evaluation index system. Using multiple machine learning methods such as SVR, GBRT and BP neural network, we first verified the effect of machine learning on the value pre-diction of GEM en-terprise, and then selected the machine learning model with the best effect to rank the importance of features. Finally, we comprehensively analyzed the factors affecting the value of GEM from the perspective of machine learning. The results show that enterprise scale, enterprise profitability, equity expansion ability, development ability and innovation ability have a great impact on the value of GEM enterprises, among which the total assets index representing enterprise scale has the greatest impact on the value of GEM enterprises.