Manual

There are many parameters to be passed to the tagging/training routine of the Erlangen-Tagger. 
To simpilify handling, every parameter can be specified in the file settings.ini as well as call by call. If a parameter is 
assigned a value both in the file settings.ini and at call time, the last value overwrites the former. The following list enumerates 
the most important parameters and gives short descriptions of them:


Training:

capClassesFile		: tag representing classes of capitalized words
contextSize		: order of the HMM
delimiters              : sentence delimiter token (full stop,colon,...)
eosTag                  : TAG for end-of-sentence token (full stop, exclamation mark,...)
numberTag               : tag for numbers
ordNumberTag            : tag for ordinal numbers
openClassesFile         : file containing a list of tags representing open word classes
morphFirst              : (yes|no) flag which specifies, if the morphology-based analyzer for unknown words is called
			  before the default prefix/suffix-based one. As the morphology API is for German language only, 
			  the default value should only be changed when working with German texts.
tagMapFile		: file which maps the tag set of the morphology API to the application tag set (German only! For other languages, a
                          pseudo file has to be provided.)
PPTrainFile		: File containing the training corpus for the rule-based post-processor
PPOutFile		: File where the learned set of rules is to be stored
taggerTrainFile		: File containing the training corpus for the HMM
taggerOutFile		: File where the learned HMM-parameters are to be stored


Tagging:

inFile			: the text to be tagged (one token per line)
paramFile		: HMM parameter file
ruleFile		: post-processing rule file
nRanks			: number of ranks of most probable tag sequences to be returned
outFile			: result file