摘要
将时间序列数据挖掘的方法应用到两国证券市场比较问题中,并在聚类分析中定义新的函数以判别最优的分类数.我们发现:在指数收盘价时间序列比较方面,中日两个证券市场的确存在一定的相似性,但中国市场的短期波动要大于日本市场.因此,如果将日本证券市场的发展历史作为中国证券市场的事件库,不足以描述和预测中国证券市场的走势.同时,在中国证券市场上,深证成指比上证综指的短期波动幅度更大,具有更多的高频噪声.
This paper applies the time series data mining method to the comparison of China＇s and Japan＇s security markets for the first time and raises a new definition to decide the optimal number of classification. We find that,there exists some similarity between Chinese market and Japanese market.However,the volatility in Chinese market is greater than in Japanese market.Thus,it will be not appropriate to take the history data of Japanese market as the event database for Chinese security market.Meanwhile,in China, Shenzhen market has a bigger volatility than Shanghai market,with higher frequency noise.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2009年第12期125-133,共9页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(70801006)
中国科学院管理
决策与信息系统重点实验室资助(70221001)
