摘要
分析了股票市场高度非线性的特点,总结了时间序列数据的分段算法,针对股票时间序列数据实现了基于斜率提取边缘点的分段表示方法,根据我国国内股票价格10%的停板限制,提出了考虑时间长度和停板规则的八元符号化表示方法。该方法既考虑了股票价格的涨幅因素,又考虑了时间长度,能有效的表示股票价格的涨跌和时间的关系,并给出理论分析和验证结果。
On the basis of analyzing high nonlinear properties of stock market and summarizing time-series data segmentation algorithm, a segmentation expression way of getting the edge point through slope is presented concerning stock time-series data, so is an eight-bit symbolization expression way taking time length and limit rules into consideration, according to the limit restriction of 10% domestic stock price , which is proved to be effective in showing the relationship between stock price change and time, followed by theoretical analysis and validated results.
出处
《电脑知识与技术》
2009年第6X期5035-5036,共2页
Computer Knowledge and Technology
关键词
时间序列
停板规则
固定窗口
符号化
行程编码
time series
limit rules
Sticky windows
symbolization
run-length coding