Trend Extraction and Similarity Matching of Financial Time Series Based on EMD Method

Data Mining of time series is a very important part of DM (Data Mining).This article proposes a trend extraction method of financial time series which are non-stationary and non-linear. And this article also gives the definition of trend of time series based on different time intervals and the criterion to measure the precision of the trend extracted. Experiments of trend extraction based on EMD method are conducted and the result verifies the effectiveness of EMD method in trend extraction of financial time series.