一种基于数据挖掘的量化投资系统的设计与实现

发布时间：2018-01-22 00:21

本文关键词： 量化投资数据挖掘金融市场交易序列计算机软件自动交易　出处：《复旦大学》2012年硕士论文　论文类型：学位论文

【摘要】：量化投资产生于1970年代末,在此后的30多年中,其得到了长足的发展。根据Bloomberg数据统计,从1998年至2008年,全球量化基金从21只发展到1184只。全球范围内,越来越多的投资者进入量化投资这个领域,量化投资已经成为一种主流的投资方式。当前,只要是对金融产品买卖投资行为的一个环节以上使用了量化技术(采用模型化、模式化、算法化、计算机程序化、自动化之一)就认为是量化投资了。因此在市场上有各种量化投资的说法,包括：量化选择金融产品、量化选择交易时间、算法交易、自动套利交易等等。相对于传统采用抽样技术建立数学模型的方法,数据挖掘技术更能反映具体金融产品交易的规律。针对金融产品投资的几个环节(研究市场和金融产品信息、研究该金融产品的历史交易、研究当前交易情况、预测未来交易、做出投资决定(买、卖、不行动)、分析交易效果),本文研究采用数据挖掘技术实现量化投资的方法,包括金融产品历史交易规律分析、交易序列模式库建设、产品交易价格的发展趋势预测、下单交易策略实现等金融产品投资主要环节的量化工作。本文的主要研发成果为：设计了一种基于数据挖掘的量化投资软件的框架,包括金融产品历史交易数据整合与挖掘、交易模式挖掘和模式库管理、产品交易价格预测与交易决定以及自动下单交易策略等内容,系统地实现了金融产品投资主要环节的量化工作。在原子交易模式的基础上,设计了TOP-K原子交易模式数据挖掘程序,设计实现了基于交易模式的金融产品聚类算法,实现了一种自动下单交易的方法,开发了一个基于数据挖掘技术的量化投资系统。
[Abstract]:Quantitative investment at the end of the 1970s, in the next 30 years, has made considerable progress. According to Bloomberg statistics, from 1998 to 2008, the global quantitative funds only from 21 to 1184. The development of the global scope, more and more investors to enter the field of quantitative investment, quantitative investment has become a mainstream investment.
At present, as long as it is more than a part of an investment behavior of financial products using quantitative techniques (using the model, model, algorithm, computer programming, automation of) that is quantitative investment. So there is a variety of quantitative investment in the market that includes quantitative selection of financial products, quantitative selection the transaction time, algorithmic trading, automatic arbitrage and so on.
Compared with the traditional method to establish the mathematical model for the sampling technology, data mining technology can reflect the specific financial products trading rules. According to several aspects of financial products investment (research market and financial product information, the financial products in the history of trading, research when trading before the situation, predict the future trading, make investment decisions (buy, sell no action), trading, analysis, this paper studies the effect) using data mining method of quantitative investment technology, including the analysis of historical financial products trading rules, trading mode construction sequence database, forecast the development trend of the product price, single quantitative financial products such as investment mainly link trading strategy.
The main research results of this paper is to design a framework of quantitative investment software based on data mining, including the financial products transaction history data integration and mining, trading pattern mining and pattern library management, product transaction price forecasting and trading decision and automatic single content trading strategy, system realizes the quantification of main part of financial products investment. Based on atomic transaction model, designed the TOP-K atomic transaction model data mining program, the design and implementation of clustering algorithm based on the mode of financial products transactions, realized an automatic method of single transaction, developed a data mining based on quantitative investment system.

【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

【相似文献】