BACK

Robust Rule-based Predictions

Introduction

This program is competitive with C4.5rules and CBA in accuracy, but is less sensitive to missing data and depends less on the default prediction. See extensive experimental results in Using Association Rules to Make Rule-based Classifiers Robust. For more theoretical background of this work, you may would like to refer to Construct robust rule sets for classification and Mining the optimal class association rule set.

This software works on Linux 9.0.

Get program

Download here

Save the file to an empty directory and then run the following commands:

(Please remove the extra .tar at the end of the downloaded file if there is one.)

tar -xzvf oac.tar.gz

chmod +x rule

chmod +x test

chmod +x translate

You will have six files in the current directory. They are anneald.data, anneald.names, anneald.test, rule, test, and translate, and the last three are executable files. 

Run program

Run the following commands:

./rule -f anneald

It will build a rule model in file anneald.robust.

Assume unseen test data are in file anneald.test.

./test -f anneald 

will give test accuracy on unseen test data.

Interpret rules

Run

./translate -f anneald

All rules in the model will be listed.

Work on your data

You need to prepare your data in C4.5 data format, as hypothyroid.names and hypothyroid.data.

If you edit your names and data files by using a window's editor, you need to use command dos2unix filenamse to convert files to UNIX format.

The program does not work on continuous attributes.

Change parameters

Run

./lirule -f hypothyroid -l 4 -s 0.1 -c 0.8

-l the maximum length of rules (default 6), -s the minimum support in a class (default 0.01), and -f the minimum confidence (default 0.5).

if you nee to build a model with more rules, run

 ./lirule -f hypothyroid -l 4 -s 0.1 -c 0.8 -r 1

 

BACK