文件列表:

dtw.h (20183, 2008-08-09)
Makefile (833, 2008-08-09)
rtreebased.h (8749, 2008-08-09)
querystrategy.h (3817, 2008-07-18)
unitesting.py (2010, 2008-07-18)
timeseries.i (287, 2008-07-18)
unittesting.cpp (1114, 2008-08-09)

LBImproved C++ Library Daniel Lemire This library comes in the form of one short C++ header file. The documentation is in the C++ comments and in this file. == Key features == 1) Fast Dynamic Time Warping nearest neighbor cost retrieval. 2) Persistence 3) External-memory: you need only a constant amount of RAM == PREREQUISITES == 1) You must first build and install the spatial index library (http://research.att.com/~marioh/spatialindex/index.html) I built this software with release 1.3.2 - May 23rd, 2008. 2) While not strictly necessary, SWIG (http://www.swig.org) is strong recommended. I interact with the library using swig and python. 3) If you are using SWIG, Python is recommended. I have used Python 2.5. == OPERATING SYSTEM == I built and ran this software with Mac OS 10.4. It also builds under Linux if you have Python 2.5 and swig installed. It should be possible to use any other Unix-like operating system, or even Windows. == BUILD == type "make" == TESTING == type "python unitesting.py" == USAGE == import dtw constraint = 0.1 n = 128 c =int(constraint*n) rtree=dtw.TimeSeriesTree("mytmpfile.bin",c,reducdim) # randomwalk(n) return a size n array for i in xrange(1000): rtree.add(randomwalk(n)) x = randomwalk(n) for mode in [rtree.LINEAR, rtree.TREE]: for algo in [ rtree.NAIVE,rtree.LB_KEOGH, rtree.LB_IMPROVED]: rtree.getNearestNeighborCost(x,algo,mode) rtree.close() # to reopen the tree, just do this: rtree=dtw.TimeSeriesTree("mytmpfile.bin")

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：