基于特征点界标过滤的时间序列模式匹配方法

动态时间弯曲距离能度量不等长的时间序列、且具有较高的匹配精度，因此广泛应用在时间序列模式匹配中。但其计算复杂度较高，制约了在大规模数据集上的应用。为了实现时间序列模式度量结果和计算复杂度的平衡，提出一种基于特征点界标过滤的时间序列模式匹配方法。首先，提出一种特征点界标过滤的特征提取方法，保留时间序列主要特征，压缩时间维度；然后，利用动态时间弯曲距离对特征序列进行相似性度量；最后，在应用数据集上对所提方法进行有效性验证。实验结果表明，所提方法在保证高精度的前提下，能有效降低计算复杂度。

Time series pattern matching method based onfeature point landmark filtering

Dynamic time warping distance can measure unequal time series and has high matching accuracy, so time series pattern matching can use it. However, its high computational complexity restricts its application in large-scale data sets. In order to balance the measurement results and computational complexity of time series pattern matching, this paper pointed out that a time series pattern matching method based on feature point landmark filtering. Firstly, it proposed a feature extraction method for feature point landmark filtering, which retained the main features of the time series and compressed the time dimension. Then, feature sequence used dynamic time warping distance for similarity measurement. Finally, the method was validated on the application data set. Experimental results show that the method can effectively reduce the computational complexity while ensuring high accuracy.