基于数据挖掘技术的创业板与主板上市企业财务指标差异研究
发布时间：2018-08-21 19:28
【摘要】：我国创业板和主板都在蓬勃发展。虽然两市场交易规则、投资者特征以及上市公司特点各有不同，但是两个市场都是符合中国国情的市场，所以两者之间具有可比性和彼此借鉴性。因此，可以找出主板与创业板两个市场企业的显著差异特征，，再基于这些差异的特征分析创业板企业的成长性特点，以及该市场的现有制度的不足，提出一些相关的建议和研究的方向，也同时对我国创业型企业的发展提供了一定的参考。 本文主要通过数据挖掘中的分类回归树，随机森林，以及Bagging算法在创业板和主板上市公司差异中的应用。首先，本文介绍了数据挖掘技术的理论和其在本文应用的优势；其次，介绍了三种挖掘算法的原理和三者之M的对比分析；同时构建了模型的指标体系，以实现对两市的差异性分类和得到相应的显著差舁特征；最后选取了2011年和2012年创业板和沪深300的上市公司，分别作为训练集和测试集，运用软件实现了上述所需的相关操作。 本文研究的结粜，对于创业板和主板来说，三个模型的对测试集的分类效果都比较理想，达到了80%以上的正确率；后续对测试集的预测正确率也很好。从而通过分类的方法实现了降维的目的，得到了三个重要分类指标。然后以主板为基准，基于这三个差异指标分析了创业板上市企业成长性特点，得到了我国创业板目前出现的一些问题。最后通过本文的研究，为后续对我同创业板的相关制度的不足和未来这方面的研究方向做了展望。
[Abstract]:China's growth Enterprise Market and the main Board are booming. Although the trading rules of the two markets, the characteristics of investors and listed companies are different, but the two markets are in line with the national conditions of China, so the two markets are comparable and can be used for reference. Therefore, we can find out the characteristics of the significant differences between the main board and the gem, and then analyze the growth characteristics of the gem enterprises based on the characteristics of these differences, as well as the shortcomings of the existing system of the market. At the same time, it provides some references for the development of entrepreneurial enterprises in China. This paper mainly applies the classification regression tree, stochastic forest and Bagging algorithm to the difference between gem and main board listed companies in data mining. First of all, this paper introduces the theory of data mining technology and its advantages in this paper. Secondly, it introduces the principle of three mining algorithms and the comparative analysis of three kinds of M, and constructs the index system of the model. Finally, the listed companies of gem and CS300 in 2011 and 2012 are selected as training set and test set, and the relevant operations mentioned above are realized by software. For the gem and the main board, the classification effect of the three models on the test set is ideal, reaching the accuracy rate of more than 80%, and the prediction accuracy rate of the test set is also very good. The aim of dimensionality reduction is realized by the method of classification, and three important classification indexes are obtained. Then taking the main board as the benchmark, this paper analyzes the growth characteristics of the gem listed enterprises based on these three indicators, and obtains some problems in the gem at present. Finally, through the research of this paper, the future research direction and the deficiency of the related system with gem are prospected.
【学位授予单位】：上海师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.51;TP311.13;F275
[Abstract]:China's growth Enterprise Market and the main Board are booming. Although the trading rules of the two markets, the characteristics of investors and listed companies are different, but the two markets are in line with the national conditions of China, so the two markets are comparable and can be used for reference. Therefore, we can find out the characteristics of the significant differences between the main board and the gem, and then analyze the growth characteristics of the gem enterprises based on the characteristics of these differences, as well as the shortcomings of the existing system of the market. At the same time, it provides some references for the development of entrepreneurial enterprises in China. This paper mainly applies the classification regression tree, stochastic forest and Bagging algorithm to the difference between gem and main board listed companies in data mining. First of all, this paper introduces the theory of data mining technology and its advantages in this paper. Secondly, it introduces the principle of three mining algorithms and the comparative analysis of three kinds of M, and constructs the index system of the model. Finally, the listed companies of gem and CS300 in 2011 and 2012 are selected as training set and test set, and the relevant operations mentioned above are realized by software. For the gem and the main board, the classification effect of the three models on the test set is ideal, reaching the accuracy rate of more than 80%, and the prediction accuracy rate of the test set is also very good. The aim of dimensionality reduction is realized by the method of classification, and three important classification indexes are obtained. Then taking the main board as the benchmark, this paper analyzes the growth characteristics of the gem listed enterprises based on these three indicators, and obtains some problems in the gem at present. Finally, through the research of this paper, the future research direction and the deficiency of the related system with gem are prospected.
【学位授予单位】：上海师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.51;TP311.13;F275
【参考文献】
相关期刊论文 前8条
1 黄文斌;金融市场中的机构投资者行为研究[J];商业研究;2004年02期
2 惠恩才;关于上市公司成长性分析[J];财经问题研究;1998年04期
3 王美今,孙建军;中国股市收益、收益波动与投资者情绪[J];经济研究;2004年10期
4 曹明;闪四清;梁海燕;;基于数据挖掘的财务预警模型设计与实现[J];计算机应用;2006年10期
5 孙红梅;邓瑶;;创业板市场上市公司成长性研究综述[J];会计之友;2011年29期
6 蔡宁,陈功道;论中小企业的成长性及其衡量[J];社会科学战线;2001年01期
7 刘e
本文编号：2196124
本文链接：https://www.wllwen.com/jingjilunwen/touziyanjiulunwen/2196124.html
