Introduction to IRLAS

IRLAS is the abbreviation for IR-Lab Lexical Analysis System. It is the basic module of IR-Lab. It includes several independent components: Atom Segmentation, Complete Segmentation, Number and Time Recognition, Unknown Word Identification and some postprocessors such as Merging of Adjoining Words, Morphologically Derived Words Recognition, New Word Identification, and so on.

The system participated in the Second International SIGHAN Word Segmentation Bakeoff which was held in July 2005. It participated in the PK open track and PK closed track and ranked 4# and 3# respectively. To see in detail the bakeoff, click here.

To test the system, click the logo above.