基于数据挖掘的套利选股模型的研究及设计

发布时间：2018-01-15 16:42

本文关键词：基于数据挖掘的套利选股模型的研究及设计　出处：《复旦大学》2012年硕士论文　论文类型：学位论文

【摘要】：随着我国计划经济向市场经济的转轨,中国证券市场在十几年的时间里迅速发展,至今沪深两个交易所已有超过两千五百家上市公司。随着股指期货,ETF等金融衍生品的出现及全流通的时代来临,中国的证券市场将进一步发展与完善。在这个越来越成熟的市场,各种先进的、创新的、定量的研究方法将有越来越大的用武之地。同时伴随着做空机制的出现,国外成熟市场的交易方式对冲、套利等新的交易手段逐步走入投资者的眼中。但对于此类交易方式的数据模型却少之又少。数据挖掘作为一项新兴的技术,最近年来,在诸多领域得到了成功的应用,主要包括：市场、销售、客户资源管理等领域。但是将数据挖掘运用于选股模型构建,投资组合盈利预测乃至将其应用于投资模型等方面的研究还不是很多。本文将数据挖掘方法引入到证券选股投资领域,重点讨论利用决策树算法对上市公司一揽子股票的阿尔法值进行数据挖掘,得出短期内阿尔法值跑赢大盘的一揽子股票,利用与期货或融资融券的对冲形成套利模型。在建立完模型以后,又采用“数据拆分”确定这些预测效果的稳健可靠。数据挖掘可以和现在已经应用较广泛的统计学和计量经济学方法相结合,以取得更好的定量分析的效果。使投资者在目前越发成熟的市场中获取稳定的投资收益。
[Abstract]:With the transition from planned economy to market economy, China's securities market has developed rapidly in more than ten years. So far, there are more than 2,500 listed companies in Shanghai and Shenzhen stock exchanges. ETF and other financial derivatives and the advent of the era of full circulation, China's securities market will be further developed and improved. In this increasingly mature market, all kinds of advanced, innovative. Quantitative research methods will be more and more useful. At the same time, with the emergence of short-selling mechanism, the trading methods of foreign mature markets will be hedged. New trading methods, such as arbitrage, have gradually come into the eye of investors. But there are few data models for such transactions. Data mining as a new technology, in recent years, has been successfully applied in many fields, including: market, sales. However, the application of data mining in stock selection model construction, portfolio profit prediction and even its application to investment model is not much. In this paper, the data mining method is introduced into the field of stock selection and investment, and the decision tree algorithm is used to mine the alpha value of the listed companies. In the short run, the alpha value outperforms the large stock market, using the hedge with futures or margin to form arbitrage model. After the establishment of the model. Data mining can be combined with the widely used statistical and econometric methods. In order to achieve better results of quantitative analysis, so that investors in the current increasingly mature market to achieve a stable return on investment.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP311.13;F832.51

【参考文献】