计算方差

Question

我有自由度和中庸之道。我需要计算标准差。

任何帮助都是非常感谢的。谢谢

Answer 1

仍然不是100%确定你要做什么，但这里有一次尝试(使用t统计量的95%范围)……我只有一个文件(这是你粘贴的3个元素)，它的名字是stats1。我正在使用awk从文件中提取相关字段，并将其抛出给R。

awk -F'>' '$1 ~/Overall/{print $2}' stats* | Rscript -e 'qt( 0.975, as.numeric (readLines ("stdin"))) '
[1]  2.940089  2.776445 12.706205

Answer 2

将一些文本处理或数据争论工具与诸如R之类的统计包结合使用当然是有意义的(特别是，您可能想要使用R的t.test(VECTOR, MU))，但是awk肯定能够胜任这项任务，甚至可能具有某些优势。无论如何，我已经附加了一个(bash)脚本，该脚本期望它的输入每行都有一个数据观察值，并按照脚本的“帮助”文档中的描述发出一行。

下面是使用t.test()从输出中提取的代码和使用脚本(t-test)运行的代码，以供比较：

$ R
> t.test(c(1,2,3), mu=0)
    One Sample t-test
t = 3.4641, df = 2, p-value = 0.07418
alternative hypothesis: true mean is not equal to 0
95 percent confidence interval:
 -0.4841377  4.4841377

$ (echo 1; echo 2; echo 3) | t-stat
3 2 0.666667 3.4641

#!/bin/bash
# USE AT YOUR OWN RISK
# For help: $0 --help
VERSION=0.1.1

# Requires: awk

BN=$(basename "$0")

function help {
cat <<EOF
Syntax: $0 [n1 [FILE]]

Compute the sample mean, the sample variance, and the t-statistic for
the array of numbers in column n1 of the input, where n1 defaults to 1,
and FILE defaults to STDIN.

The output consists of a single line of four values:

n mean variance t

where "n" is the sample size (possibly 0),
      "mean" is the sample mean,
      "variance" is the average of the squared deviations from the mean, and
      "t" is the sample t-statistic, computed as 
           mean * sqrt( n * ss / (n-1));

If a statistic cannot be computed, an asterisk (*) is used in its place.

If the value in the specified field in the input is null or
non-numeric, it is ignored.

If a file is specified as "-" then STDIN is read.

Example:

$BN <<< $'1\n2\n3'
3 2 2 3.4641

EOF
}

function die     { echo "$BN: $@" >&2 ; exit 1 ; }

case "$1" in
    -h | --help )
    help
    exit
    ;;
    -v | --verbose )
    VERBOSE=1
    shift
    ;;
esac

# Default:

file=-

if [ $# = 0 ] ; then
    f1=1
elif [ "$1" = - ] ; then
    f1=1
else
   f1="$1"
   shift
   check=`expr "$f1" : "\([0-9]*\)"`
   if [ "$f1" != "$check" ] ; then die a non-numeric field was specified ; fi
fi

if [ $# != 0 ] ; then file="$1" ; fi

awk -v verbose="$VERBOSE" -v f1="$f1" '
  BEGIN { a1=0; s1=0; n=0; }
  NF < f1 { next }
  $f1 == "" || ($f1+0 != $f1) { next }
  { n++; a1 += $f1; s1 += $f1*$f1; }

  END {
    s="*";
    if ( n == 0 ) {print 0, s, s, s; exit }
    avg = a1/n;
    var = s1/n - avg*avg;
    if (var == 0) { print n, avg, var, s; exit}
    tstat = avg * sqrt( (n-1) / var);
    print n, avg, var, tstat; 
  } ' "$file"