基于金融时间序列的符号聚合近似测度的改进

Improvement of Approximate Measure of Symbol Aggregation Based on Financial Time Series

针对金融时间子序列模式匹配准确率低的问题,提出了一种在金融时间序列中定位图形模式的新方法,即改进的符号聚合近似表示方法。利用子序列每个区段的最小值、平均值和最大值进行符号化,并采用余弦相似度作为模式相似性的衡量标准。采用香港股票市场的恒生指数HIS的历史数据进行仿真实验,实验结果表明,无论是从匹配时间还是匹配的准确性角度分析,所提算法均比传统的SAX方法和TD_SAX方法更有效,所需时间较短且可以产生更少的假阴性和假阳性序列。

Aiming at the low accuracy of pattern matching in financial time series,this paper proposes a new method of locating pattern in financial time series,which is enhanced symbolic aggregate approximation vector space.The minimum,average,and maximum values for each segment of the subsequence are used for symbolization,and cosine similarity is used as a measure of pattern similarity.Using the historical data from the HSI of the Hong Kong stock market to conduct a simulation experiment,the experimental results show that the proposed algorithm is more efficient than traditional SAX algorithm and TD_SAX algorithm in terms of matching time and matching accuracy.Less time is required and fewer false-negative and false-positive sequences are generated compared with SAX and TD_SAX.