Winnow (original) (raw)

Implements Winnow and Balanced Winnow algorithms by Littlestone.

For more information, see

N. Littlestone (1988). Learning quickly when irrelevant attributes are abound: A new linear threshold algorithm. Machine Learning. 2:285-318.

N. Littlestone (1989). Mistake bounds and logarithmic linear-threshold learning algorithms. University of California, Santa Cruz.

Does classification for problems with nominal attributes (which it converts into binary attributes).

BibTeX:

@article{Littlestone1988, author = {N. Littlestone}, journal = {Machine Learning}, pages = {285-318}, title = {Learning quickly when irrelevant attributes are abound: A new linear threshold algorithm}, volume = {2}, year = {1988} }

@techreport{Littlestone1989, address = {University of California, Santa Cruz}, author = {N. Littlestone}, institution = {University of California}, note = {Technical Report UCSC-CRL-89-11}, title = {Mistake bounds and logarithmic linear-threshold learning algorithms}, year = {1989} }

Valid options are:

-L Use the baLanced version (default false)

-I The number of iterations to be performed. (default 1)

-A Promotion coefficient alpha. (default 2.0)

-B Demotion coefficient beta. (default 0.5)

-H Prediction threshold. (default -1.0 == number of attributes)

-W Starting weights. (default 2.0)

-S Default random seed. (default 1)