ePCL
Version 3.0
============


Emerging patterns are itemsets whose frequencies change sharply from one class to the other. They
are thus useful for classification problems. PCL is an example of efficient classification algorithms which
leverage the prediction power of emerging patterns. It first selects the top-K emerging patterns of each class
that match a testing instance, and then uses these selected patterns to decide the class label of the testing
instance. ePCL is a variation of PCL that incorporate the Central Limit Theorem and pattern space maintenance
algorithms ro efficiently determine the best value of K for PCL, which can improve the accuracy of PCL greatly.


Usage:

	mmm3 [datafile] [topK] [number of fold] [maxK] [#Runs]


Example: mmm3 zoo.dat 10 10 50 50

- topk=10 (if topk=-1, parameter setting will be implemented)
- 10 folds cross validation is used
- all values of K from 1 to 50 will be examined
- 50 iterations will be performed


short form: mmm3 [datafile] [topK]

by default: number of fold = 10
	    maxk = 50
	    #Runs=50
 	


Credits:

The EPCL program is written by Thanh-Son Ngoh (dcsnts@nus.edu.sg). 
The work was supported by a A*STAR SERC PSF grant SERC 072 101 0016.


If you use this software, please cite the following papers:

Thanh-Son Ngo, Mengling Feng, Guimei Liu, Limsoon Wong.
Efficiently Finding the Best Parameter for the Emerging Pattern-based Classifier PCL.
Manuscript, 6 December 2009.