id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_37btiicl5fhs5lb2yfxzlmafnu	Daniel Gordon	Local-shapelets for fast classification of spectrographic measurements	2015	17	.pdf	application/pdf	8535	716	67	A shapelet is a subsequence extracted from one of the time series in the data set. all the time series in the training set, and for each shapelet calculating its distance to each time first optimizes the distance calculation of a shapelet to a time series. The Euclidean distance between a time series T of length m and a local-shapelet presents two time series from the same class of the data set Lightning7. Figure 2: Two examples from different classes of the Lightning7 data set illustrating the need for a tolerance range. Our goal is to show that for data sets from the field of spectroscopy, the usage of localshapelets reduces the time complexity during both training and classification phases in comparison with global-shapelets (i.e., non-local shapelets) without degrading accuracy. of examples in the training and test sets, the number of classes, the length of the time series and	./cache/work_37btiicl5fhs5lb2yfxzlmafnu.pdf	./txt/work_37btiicl5fhs5lb2yfxzlmafnu.txt