基于离散余弦变换的时间序列相似性检索

Time Series Similarity Based on Discrete Cosine Transform

在时间序列相似性研究领域已经发展了多种方法用于时间序列的表示,以达到降低序列维度的目的.作为一种经典的时域-频域转换方法,离散余弦变换目前已经在图形图像处理等领域得到了广泛的应用.将此方法应用于时间序列的表示上,在变换后的数据上进行相似性查询等操作.实验表明,相对以前的方法,这种方法具有明显的性能提升.

In the research domain of time series similarity, there have been many approach that are developed to represent time series in the order of reducing the dimensionality. As one of the classic time domain-frequency domain transformation method, the discrete cosine transform has been widely used in the field of graphics and image processing. In this paper, we used this method to represent time series, and implenmet querying operation etc. on ttie transformed data. Our experiments indicate that this method obviously improve the performance compared with previous method..