Similarity Search on Financial Time Series based on DTW and NMF
Euclidean distance measure or its transformatoins are often utilized for similarity measure, which can be well suitable for dealing with non-time-series data. However, Euclidean distance measure is not a good method to the time series that is warped non-linearly by stretching or shrinking along its time-axis, which often leads to an inaccurate result. In order to solve this problem, this paper puts forward an effective measure technique based on Non-negative Matrix Factorization (NMF) and Dynamic Time Warping (DTW). The measure technique is described as follows: Applying NMF many times to reduce dimensionality of time series to obtain varied dimensions time series. Then beginning similarity search with DTW. At last, we can easily find similar time series will cluster through comparing results of measures.
Time series Similarity measure DWT NMF
Zunxiong Liu Tianqing Zhou
East China Jiaotong University
国际会议
武汉
英文
1112-1116
2010-06-06（万方平台首次上网日期，不代表论文的发表时间）