Nbins

Data Mining
CS 510 (DM)
Winter,2004
home | news | site map
review | project | subject | group
weka | mining | gawk | bash
modeling | reference | pods
Display: big | small

Why all the scripting?
 copyleft() {
        cat<<-EOF
        nbins: discretize numerics into n bins
        Copyright (C) 2004 Tim Menzies
        This program is free software; you can redistribute it and/or
        modify it under the terms of the GNU General Public License
        as published by the Free Software Foundation, version 2.
        This program is distributed in the hope that it will be useful,
        but WITHOUT ANY WARRANTY; without even the implied warranty of
        MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
        GNU General Public License for more details.
        You should have received a copy of the GNU General Public License
        along with this program; if not, write to the Free Software
        Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
        EOF
  }

Motivation

Some learners can only handle discrete (a.k.a. nominal or non-numeric) variables. Here's a simple discretizer that finds all numbers in a table of data and discretizes them into bins of size max-min/n.

Of course, this is just the tip-of-the-iceberg in handling the discretization problem. For a more complete analysis, see [Do95]. In fact, this code is just a demonstrator for my data mining students so they can try different discretization methods.

[TOP]


Usage

 usage() {
        cat<<-EOF
        Usage: nbins  [FLAGs] FILE
        Simple (very) discretization of all numeric columns in FILE into n bins
        Flags: 
          -h             print this help text
          -l             copyright notice
          -m CHAR        missing value character; default="$Q$missing0$Q
          -n NUM         number of bins; default=$Q$bins0$Q
          -s CHAR        field separator; default=$Q$sep0$Q
          -x            run an example
        EOF
        exit
 }

Examples

Here's the distribution of numbers in the last column of the file nbinseg.dat:

 sirius:~/public_html/dm [769]$ bars -r 2 nbinseg.dat 
    10|   3| ***
    12|  10| **********
    14|  40| ****************************************
    16|  39| ***************************************
    18|  39| ***************************************
    20|  40| ****************************************
    22|  26| **************************
    24|  32| ********************************
    26|  36| ************************************
    28|  28| ****************************
    30|  24| ************************
    32|  27| ***************************
    34|  18| ******************
    36|  13| *************
    38|  11| ***********
    40|   5| *****
    42|   1| *
    44|   5| *****
    46|   1| *

nbins can bunch these up into (e.g.) 5 bins:

 sirius:~/public_html/dm [770]$ nbins -n5 nbinseg.dat | bars -r 2 
    8|   7| **
   16| 124| **************************************
   22| 129| ****************************************
   30|  92| *****************************
   38|  39| ************
   46|   7| **

Installation

First, if you have installed anything from this site before, save your config file to somewhere safe.

Second, copy the following files to your directory (from either ~timm/public_html/dm or http://www.cs.pdx.edu/~timm/dm or from http://www.cs.pdx.edu/~timm/dm/nbins.zip):

Standard files
config, and nbins.

Support code
nbins.awk and lib.awk

Example files
nbinseg.dat

Third, make nbins executable:

 chmod +x nbins

Fourth, compare your safe version of config with the new version you just copied and fix up any paths.

Five, edit this file and config. The first line of this file should point to your local bash shell. and you'll need to check at least the #paths section in config

Check that all it works:

 nbins -x

If the installation worked, then you should see a file returned with a smaller range of numbers in each column.

[TOP]


Source code

Settings

Defaults:

 sep0=","
 missing0="?"
 bins0=10

Paths:

 . config

Minor details:

 Q="\""

Demo code

 nbinsDemo() { 
        main nbinseg.dat
        }

Main

 main() {
  $gawk -F"$sep" -f lib.awk -f nbins.awk \
        Bins=$bins Missing="$missing"  \
        pass=1 $1 pass=2 $1
 }

nbins.awk: The Worker

 BEGIN {
   FS=OFS=",";
   Missing="?";
   Bins=10;
   Max; Min;     #arrays of min/max numbers in each column
 }
 FNR==1 && pass==1 { for(I=1;I<=NF;I++) { 
                        Max[I]= -1 * Inf;
                        Min[I]= Inf};
 }
 pass==1 {
   for(I=1;I<=NF;I++) {
     if ($I == "?") continue;
     if (! number($I) ) {
       Symbol[I];
       continue;}
     else {
       Max[I]=most(Max[I],$I);
       Min[I]=least(Min[I],$I);
       if (novel(I,$I,Seen)) Range[I]++
 }}}
 pass==2 {
   for(I=1;I<=NF;I++) $I= discretize(I,$I);
   print $0;
 }
 function discretize(col,i,  r) {
   if ( Symbol[col] )     return i;
   if ($I == Missing)     return i;
   if (Range[col] < Bins) return i;
   r=(Max[col]- Min[col])/Bins;
   return  round(i/r)*r;
 }

Command line processing

 demo=""
 while getopts "hln:s:x" flag
 do case "$flag" in
        h)  usage; exit ;;
        l)  copyleft; exit;;
        m)  missing=$OPTARG;;
        n)  bins=$OPTARG;;
        s)  sep=$OPTARG;;       
        x)  demo="nbinsDemo nbinseg.dat";;
    esac
 done
 shift $(($OPTIND - 1))
 missing=${missing:=$missing0}
 bins=${bins:=$bins0}  
 sep=${sep:=$sep0}  
 if [ -n "$demo" ]
 then $demo
      exit
 else  main $1
 fi

[TOP]


References

Do95
Supervised and Unsupervised Discretization of Continuous Features (1995) by James Dougherty and Ron Kohavi and Mehran Sahami, ICML, pages 194-202, 1995, http://www.cs.pdx.edu/~timm/dm/dougherty95supervised.pdf

[TOP]


Credits

Author

Tim Menzies , tim@menzies.us, http://menzies.us

Software

This page generated by Site: see http://www.cs.pdx.edu/~timm/dm/site.html

Acknowledgements

This site is built using PerlPod.

Style sheet switching method taken from Eddie Traversa's excellent and simple-to-apply tutorial: http://dhtmlnirvana.com/content/styleswitch/styleswitch1.html.

Search engine powered by ATOMZ http://www.atomz.com/search/. Note, the indexes to this site are only updated weekly (heh, its a free service- what more ja want?).

Icons on this site come from http://www.sql-news.de/rubriken/olap.asp and http://www.ifnet.it/webif/centrodi/eng/toolbar.htm.

The JAVA machine learners used at this site come from the extensive data mining libraries found in the University of Waikato's Environment for Knowledge Analysis (the WEKA) http://www.cs.waikato.ac.nz/ml/weka/

[TOP]


Legal

Copyright

Copyright (C) Tim Menzies 2004

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 2; see http://www.gnu.org/copyleft/gpl.html. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.

Disclaimer

The content from or through this web page are provided 'as is' and the author makes no warranties or representations regarding the accuracy or completeness of the information. Your use of this web page and information is at your own risk. You assume full responsibility and risk of loss resulting from the use of this web page or information. If your use of materials from this page results in the need for servicing, repair or correction of equipment, you assume any costs thereof. Follow all external links at your own risk and liability.

[TOP]