文件列表:

hist\.cvsignore (16, 2006-10-21)
hist\.tkenvrc (1585, 2010-01-08)
hist\ChangeLog (2143, 2006-10-21)
hist\hist.cpp (4051, 2006-10-21)
hist\hist.dsp (7780, 2009-07-26)
hist\hist.dsp.in (7662, 2006-10-21)
hist\hist.dsw (533, 2006-10-21)
hist\hist.exe (1634304, 2006-10-21)
hist\hist.ned (1351, 2006-10-21)
hist\Makefile.vc (3312, 2006-10-21)
hist\omnetpp.ini (559, 2006-10-21)
hist (0, 2009-07-26)

Histogram demo ============== A demo for the density estimation/histogram classes. Demonstrates: - collecting observations into statistics objects - saving statistics objects to file and restoring them About the model The network consists of a single module. It goes into a loop to generate random numbers. The numbers are inserted into histograms and other density estimation objects using different algorithms. Some of the objects use the first n numbers to estimate a range, and set up histogram cells accordingly. When you start the model, it will ask a couple of questions; it is OK to accept the defaults offered. When you run the model, several inspector windows will appear which display histogram-like diagrams. (The corresponding objects you can find in the object tree on the left side of the main window, after opening the "Histograms" and "local-objects" nodes.) The following classes are demonstrated: * cDoubleHistogram A histogram with equidistant cells. Cell boundaries are real numbers (doubles). In the default scenario of this demo, the lower histogram bound is zero, and the upper bound is estimated from the first n (by default 50) observations. * cLongHistogram A histogram with equidistant cells. Cell boundaries are integers. Histogram bounds are set up similarly to cDoubleHistogram. * cVarHistogram A histogram with arbitrary cells. Cell boundaries can be set up manually (via the addBinBound() member function of the C++ class), but it is also possible to have them set them up automatically based on the first n observations such that each cell contains approximately the same number of observations. (Cell boundaries don't change afterwards.) The latter method is used in this demo. * cPSquare Implementation of the P^2 (P-square) algorithm from Jain and Chlamtac. It is an on-line density estimation method that works by estimating and keeping track of quantiles. The resulting (calculated) histogram will thus contain roughly equi-probable cells. * cKSplit An adaptive histogram structure which splits those cells that receive "too many" observations (this is defined by a split criterion), thus "refining the resolution" at (or "zooming in" on) the interesting parts of the distribution.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：