org.antlr.runtime.misc
Class Stats

java.lang.Object
  org.antlr.runtime.misc.Stats

public class Stats
extends java.lang.Object

Stats routines needed by profiler etc... // note that these routines return 0.0 if no values exist in the X[] // which is not "correct", but it is useful so I don't generate NaN // in my output

Constructor Summary

Stats()

Method Summary

static double	avg(int[] X) Compute the sample mean
static java.lang.String	getAbsoluteFileName(java.lang.String filename)
static int	max(int[] X)
static int	min(int[] X)
static double	stddev(int[] X) Compute the sample (unbiased estimator) standard deviation following: Computing Deviations: Standard Accuracy Tony F.
static int	sum(int[] X)
static void	writeReport(java.lang.String filename, java.lang.String data)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Stats

public Stats()

Method Detail

stddev

public static double stddev(int[] X)

Compute the sample (unbiased estimator) standard deviation following: Computing Deviations: Standard Accuracy Tony F. Chan and John Gregg Lewis Stanford University Communications of ACM September 1979 of Volume 22 the ACM Number 9 The "two-pass" method from the paper; supposed to have better numerical properties than the textbook summation/sqrt. To me this looks like the textbook method, but I ain't no numerical methods guy.

avg

public static double avg(int[] X)

Compute the sample mean

min

public static int min(int[] X)

max

public static int max(int[] X)

sum

public static int sum(int[] X)

writeReport

public static void writeReport(java.lang.String filename,
                               java.lang.String data)

getAbsoluteFileName

public static java.lang.String getAbsoluteFileName(java.lang.String filename)