This is the code for Accelerated Higher Order Logistic Regression.
This code is supporting the following research paper:
ALRn: accelerated higher-order logistic regression, Zaidi, N.A., Webb, G.I., Carman, M.J., Petitjean, F. and Cerquides, J. Machine Learning (2016). doi:10.1007/s10994-016-5574-8.
When using this repository, please cite:
@article{Zaidi2016,
title={ALRn: accelerated higher-order logistic regression},
author={Zaidi, Nayyar A and Webb, Geoffrey I and Carman, Mark J and Petitjean, Fran{\c{c}}ois and Cerquides, Jes{\'u}s},
journal={Machine Learning},
volume={104},
number={2-3},
pages={151--194},
year={2016},
publisher={Springer}
}
An Example command line of the code:
java ALR.BVDcrossvalx -t /Users/nayyar/WData/datasets_DM/nursery.arff -i 2 -x 2 -W ALR.wdAnJE -- -S "A1JE" -P "dCCBN" -I "Flat" -O "GD"
BVDcrossvalx does 'i' rounds of 'x' fold cross validation and calls ALR classifier.
ALR takes in following arguments:
-S A1JE, A2JE, A3JE, A4JE, A5JE
This specifies to use linear, quadratic, cubic, quartic or quintic features
-P MAP, dCCBN, wCCBN
This specifies to use either generative MAP learning or LR or ALR
-I Flat, Indexed, IndexedBig, BitMap
This depends on the data and its dimensions and implements the model effectively