Enhancing Natural Language Querying using Ontolection Trainer

Ontolection Trainer is a command line utility to train machine learning algorithms on a text corpus in order to output an ontolection in Watson™ Explorer Engine format for use in query expansion.

Ontolection Trainer uses the PEAR files that ship with Watson Explorer in order to support all languages that are officially supported by Watson Explorer. It requires these files in order to run. See Lexical Analysis Streams for the list of languages.

Training on text corpora that contain text in multiple different languages is not supported.

Ontolection Trainer does not ship with any example text corpora; the user must provide a text corpus on which to train. For an example of how to generate a text corpus from a Watson Explorer search collection, see the example Ontolection Trainer helper scripts. They are located in the examples/nlq/ontolectiontrainer directory within the Watson Explorer Engine install directory.