nnlim

来自「it is the Data Mining Algorithm source c」· 代码 · 共 127 行

#!/bin/bash

sel=${1:-all}
limit=${2:-1e-6}
cnt=${3:-20}
off=${4:-1}
inc=${5:-7}
add=$6

explist="1.05 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0"
mmtlist="0.05 0.1 0.15 0.2 0.25 0.3 0.4 0.5 0.6 0.7 0.8 0.9"

#-----------------------------------------------------------------------

function run ()
{                               # --- run one clustering trial
  mode=${1:-none}               # get the training mode
  for (( i = 0; i < cnt; i++ )); do
    cli -S$(( (inc*i)+1 )) -A$mode $2 $args 2> $data.2
    e=`sed -n "s/.*\[\([0-9]*\) epoch(s)\].*/\1/p" $data.2`
    echo $e >> $data.1
    if (( e > 1000 )); then break; fi;
  done
  gawk -v data=$data -v ncls=$ncls -v mode=$mode -v opts="($2)" '
    BEGIN { e = s = n = 0; }
    { e += $1; s += $1*$1; n++; }
    END {
      printf("%-7s %1d %-9s %-8s ", data, ncls, mode, opts);
      if (n <= 0) n = 1;
      s = sqrt((s -e*e/n)/n); e /= n; 
      printf(": %6.1f (~%6.1f)\n", e, s);
    }' $data.1 | tee -a nnlim.res
  rm -f $data.cls $data.1 $data.2
}  # run ()

#-----------------------------------------------------------------------

function modes ()
{                               # --- try different training modes
  data=${1:-iris}               # get the data name
  ncls=${2:-2}                  # get the number of clusters
  args="-$3c$ncls -e5000 -T$limit $add $data.dom $data.tab $data.cls"
  run none                      # start with a normal run
  if [[ $4 != "" ]]; then explist="$4"; fi
  for x in $explist; do         # try different expansion factors
    run expand   -g$x; done
  if [[ $5 != "" ]]; then mmtlist="$5"; fi
  for x in $mmtlist; do         # try different momentum terms
    run momentum -m$x; done
  run adaptive                  # adaptive expansion factor
  run resilient                 # resilient backpropagation
  run quick -g2                 # quick backpropagation
}  # modes ()

#-----------------------------------------------------------------------

function all ()
{
  echo shape: $1 | tee -a nnlim.res
  modes iris    3 $1
  modes wsel    3 $1
  modes wsel    6 $1
  modes wine    3 $1
  modes wine    6 $1
  modes breast  2 $1
  modes abalone 3 $1
  modes abalone 6 $1
}  # all ()

#-----------------------------------------------------------------------

echo limit: $limit > nnlim.res

if [[ $sel == "all" ]]; then
  all           # fuzzy c-means algorithm
  all v         # axes-parallel Gustafson-Kessel algorithm
  all V         # normal        Gustafson-Kessel algorithm
fi

if [[ $sel == "best" || $sel == "iris" || $sel == "iris3" ]]; then
  modes iris 3 "" 1.4 0.25
  modes iris 3 v  1.4 0.3
  modes iris 3 V  1.8 0.5
fi

if [[ $sel == "best" || $sel == "wsel3" ]]; then
  modes wsel 3 "" 1.6 0.3
  modes wsel 3 v  1.5 0.2
  modes wsel 3 V  1.6 0.3
fi

if [[ $sel == "best" || $sel == "wsel6" ]]; then
  modes wsel 6 "" 1.8 0.6
  modes wsel 6 v  1.9 0.7
  modes wsel 6 V  2.0 0.9
fi

if [[ $sel == "best" || $sel == "wine3" ]]; then
  modes wine 3 "" 1.4 0.2
  modes wine 3 v  1.5 0.2
  modes wine 3 V  1.9 0.7
fi

if [[ $sel == "best" || $sel == "wine6" ]]; then
  modes wine 6 "" 1.9 0.6
  modes wine 6 v  1.9 0.7
  modes wine 6 V  1.9 0.6
fi

if [[ $sel == "best" || $sel == "breast" || $sel == "breast2" ]]; then
  modes breast 2 "" 1.2 0.05
  modes breast 2 v  1.5 0.2
  modes breast 2 V  1.8 0.5
fi

if [[ $sel == "best" || $sel == "abalone3" ]]; then
  modes abalone 3 "" 1.6 0.3
  modes abalone 3 v  1.6 0.3
  modes abalone 3 V  1.9 0.6
fi

if [[ $sel == "best" || $sel == "abalone6" ]]; then
  modes abalone 6 "" 1.9 0.9
  modes abalone 6 v  1.8 0.5
  modes abalone 6 V  1.9 0.5
fi