基于数据挖掘的数量化模型选股分析平台

发布时间：2018-02-15 07:32

本文关键词： 数据挖掘分类算法数据库模型选股股票分析　出处：《电子科技大学》2012年硕士论文　论文类型：学位论文

【摘要】：伴随着计算机、信息技术的飞速发展，信息数据的存储和获取技术也得到了前所未有的发展，各个行业领域都产生了大量数据，而从这些数据中怎样提取到对我们有用的数据，仅仅按照常规的方法已很难去解决，而近年来产生的数据挖掘则技术则可以发现那些隐藏在海量数据的、具有一定规律的、对我们有用的信息数据，所谓数据挖掘是一种应用模型发现知识、提取有用数据的过程，我们可以用这些数据进行分析和预测。股票市场是我国市场经济不可缺少的组成部分，在经济发展中起着不可代替的作用，如何能够比较正确的分析和预测股票未来的走势，对金融投资方面来说有着非常重要的意义。但是股票的价格走向是受很多因素的影响，所以说炒股是一个有着非常不确定性的复杂过程。对它建立某种固模型是有一定困难的，同时股票相关的数据越来越庞大，而这些数据中常常包含着股票价格走势的规律性。而近些年来新兴发展起来的据挖掘技术则是一种可以满足从这种海量数据之中，获取有价值的数据的新的数据处理技术，因而如何对股票数据利用挖掘技术进行分析和处理，并做出趋势预测具有重大的理论和实际的意义。本论文主要讨论了基于现金流的选股模型，将其理论数量化抽象生成一套关于选择股票的公式分类规则，并对选择的样本数据进行预处理，转换构造挖掘所需要指标和属性，然后利用数据挖掘技术中的公式分类技术和数据库中的一些聚集函数，结合SQL语句对股票数据进行分析、预测，并对挖掘结果进行了必要的检验。根据实际结果证明从模型获得的公式分类算法进行选股是可行的。用户可以根据这类规则快速的选择出有投资价值的个股，然后进一步的去分析和预测或直接投资。最后根据所论证挖掘模型以及分类所用算法设计并开发了一个实际的模型分析选股平台，并对其功能和性能等进行了必要的测试，它可以对股票数据进行多维的分析预测，作为投资者的投资决策的辅助工具，是利用数据挖掘技术结合华尔街著名选股模型理论，分析大量与股票相关的信息数据，，并做出未来走势预测，具有一定实用意义。
[Abstract]:With the rapid development of computer and information technology, the storage and acquisition technology of information data has been developed unprecedented. A large number of data have been produced in various fields, and how to extract the useful data from these data. It is difficult to solve only according to the conventional method, but the data mining technology produced in recent years can find the information data which is hidden in the massive data, has certain regularity, and is useful to us. Data mining is a process of applying model to discover knowledge and extract useful data, which can be used for analysis and prediction. The stock market is an indispensable part of our market economy and plays an irreplaceable role in economic development. How can we correctly analyze and predict the future trend of the stock market? It is very important for financial investment. But the price trend of stock is influenced by many factors, so stock speculation is a complicated process with very uncertainty. It is difficult to build some kind of solid model for it. At the same time, stock data are getting bigger and larger, and they often contain the regularity of stock price movements. And the newly developed data mining technology in recent years is a way to satisfy this huge amount of data. New data processing technology for obtaining valuable data, so how to analyze and process stock data using mining technology and make trend prediction has great theoretical and practical significance. This paper mainly discusses the stock selection model based on cash flow, and abstracts the theoretical quantification to generate a set of formula classification rules about stock selection, and preprocesses the selected sample data, and transforms and constructs the indexes and attributes needed for mining. Then using the formula classification technology in the data mining technology and some aggregation functions in the database, combined with the SQL statement to analyze and predict the stock data, According to the actual results, it is proved that it is feasible to select stocks by the formula classification algorithm obtained from the model. Users can quickly select the stocks with investment value according to this kind of rules. Then further analysis and prediction or direct investment. Finally, a practical model analysis and stock selection platform is designed and developed according to the demonstrated mining model and the algorithm used in classification, and its function and performance are tested, which can be used for multidimensional analysis and prediction of stock data. As an assistant tool for investors' investment decision, it is of practical significance to use data mining technology combined with Wall Street famous stock selection model theory to analyze a large number of information data related to stocks and to predict the future trend.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP311.13;F832.51

【参考文献】