摘要
提出了一种从股价时间序列中提取形态特征的股价研究方案。文中使用该系统对真实股市进行了仿真实验,并取得了很好的结果。系统可以使大家从在大量股票图线中寻找重要技术形态的繁重工作中解脱出来,有针对性地对出现这些形态的股票进行更为深入的指标技术分析和基本面分析,从而提供更为完备、更为准确的投资策略。对于与股市波动直接利益相关的个人投资者而言,自动形态分析系统为他们提供专家级的第一手信息,帮助他们捕捉最佳的买卖时机。研究的算法可以在短时间内从海量的股价数据中搜索大量的形态样本,为相关研究人员提供帮助。
In this paper,offer a research scheme for the forecasting of share price trends based on sentiment orientation of financial information. At the present age, the impact, which caused by Internet financial information, can＇t be neglected increasingly on financial market. In the face of massive information,most of which are unstructured text data,adopt the current sentiment classification algorithm and then add our value of sentiment orientation to timeseries model as external variable. At last,give the forecast of share price＇s trends and fluctuation. The result shows the relativity between financial information in Intemet and the two attributes of share price which prove the scheme is effective.
出处
《计算机技术与发展》
2009年第10期1-5,10,共6页
Computer Technology and Development
基金
国家自然科学基金(70571003)
国家863高技术研究发展计划投资项目(2007AA01Z437)
关键词
计算机仿真
股价趋向预测
股市波动
computer simulation
forecasting of share price trends
fluctuation of share price
