public static void main(String[] args) throws Exception {
// Parse command line flags and arguments
Map<String, String> argMap = CommandLineUtils.simpleCommandLineParser(args);
// Set up default parameters and settings
String basePath = ".";
boolean verbose = false;
// Update defaults using command line specifications
// The path to the assignment data
if (argMap.containsKey("-path")) {
basePath = argMap.get("-path");
}
System.out.println("Using base path: " + basePath);
// Whether or not to print the individual errors.
if (argMap.containsKey("-verbose")) {
verbose = true;
}
// Read in data
System.out.print("Loading training sentences...");
List<TaggedSentence> trainTaggedSentences =
readTaggedSentences(basePath + "/en-wsj-train.pos", true);
Set<String> trainingVocabulary = extractVocabulary(trainTaggedSentences);
System.out.println("done.");
System.out.print("Loading in-domain dev sentences...");
List<TaggedSentence> devInTaggedSentences =
readTaggedSentences(basePath + "/en-wsj-dev.pos", true);
System.out.println("done.");
System.out.print("Loading out-of-domain dev sentences...");
List<TaggedSentence> devOutTaggedSentences =
readTaggedSentences(basePath + "/en-web-weblogs-dev.pos", true);
System.out.println("done.");
System.out.print("Loading out-of-domain blind test sentences...");
List<TaggedSentence> testSentences =
readTaggedSentences(basePath + "/en-web-test.blind", false);
System.out.println("done.");
// Construct tagger components
// TODO : improve on the MostFrequentTagScorer
LocalTrigramScorer localTrigramScorer = new MostFrequentTagScorer(false);
// TODO : improve on the GreedyDecoder
TrellisDecoder<State> trellisDecoder = new GreedyDecoder<State>();
// Train tagger
POSTagger posTagger = new POSTagger(localTrigramScorer, trellisDecoder);
posTagger.train(trainTaggedSentences);
// Optionally tune hyperparameters on dev data
posTagger.validate(devInTaggedSentences);
// Test tagger
System.out.println("Evaluating on in-domain data:.");
evaluateTagger(posTagger, devInTaggedSentences, trainingVocabulary, verbose);
System.out.println("Evaluating on out-of-domain data:.");
evaluateTagger(posTagger, devOutTaggedSentences, trainingVocabulary, verbose);
labelTestSet(posTagger, testSentences, basePath + "/en-web-test.tagged");
}
