Package edu.stanford.nlp.ie.pnp

Character-based classifier and generator for proper noun phrases (PNPs).

See:
Description

Class Summary

ConfusionMatrixPlotter	Visualization utility for confusion matrices.
DataGenerator	Creates training/test/answer files from input files.
DataGenerator.Example	Stores a category number and an example text.
HumanClassifier	Command-line PNP test to measure human accuracy.
PnpClassifier	Statistical classifier of unseen proper noun phrases.
PnpGenerator	Generates novel PNPs using a trained PnpClassifier.
TestScriptGenerator	Generates a shell script file to run batch tests for PnpClassifier.

Package edu.stanford.nlp.ie.pnp Description

Character-based classifier and generator for proper noun phrases (PNPs). PNPs are essentially named-entities (e.g. company names, people names, place names, etc.) PnpClassifier learns a generative model of PNPs from training data (lists of PNPs for each category), which can be used both to classify unknown PNPs and to generate novel PNPs. PnpGenerator uses a PnpClassifier to generate new PNPs after being trained on some example. The other files are for generating data splits and test files, for assessing the accuracy of humans at the task of PNP classification, and for visualizing confusion matrices.

The data files used for PnpClassifier are on the nlp machines in /u/nlp/data/pnp.