摘要
本文将模式识别分类技术应用于股票市场进行搜索选股。本项目采用股票价格走势来构造特征空间,依据样本股票价格走势类型对股票市场上的所有股票进行模式识别分类,并将所获得的分类结果中的第一类股票作为搜索选股的目标股票。通过对沪深两市股票进行的实证研究分析,结果表明:采用模式识别分类技术依据股票价格走势进行搜索选股是可行的,具有良好的实时性和较高的针对性,实用性较强,并可进一步改进。
This article presents a new method of search for stock picking by applying pattern recognition classification techniques on the stock market.The project method uses the stock price movements to construct the feature vector space model.All stocks on the stock market are classified according to the sample type of stock price movements and the classification results obtained by the first category of stock will be as target stocks after the search for stock picking.Through empirical research and analysis on stocks on the Shanghai and Shenzhen market,the results show that it is feasible that using pattern recognition classification techniques based on stock price movements to search for stock selection,with good real-time,high purposeful,usefulness strong and can be further improved.
出处
《微计算机信息》
2011年第1期281-284,共4页
Control & Automation
基金
基金申请人:王利
项目名称:上海市研究生创新基金项目
基金颁发部门:上海市教育委员会(JWCXSL0902)
关键词
K-近邻法
股价曲线
模式识别分类
K-Nearest Neighbor Classifier
Stock Price Curve
Pattern Recognition Classification
