4ft-AR2NL module

Author(s):

Petr Strossa, Zdeněk Černý

Responsibility:

Petr Strossa (theory), Zdeněk Černý (software), ? (help)

Description:

4ft-AR2NL system converts association rules (AR – Association Rules) into natural language (NL – Natural Language). Let us suppose that we have got an association rule (from LISp-Miner system) in this form:

Activity in a Job(he mainly sits) ∧ Activity after a Job(he mainly sits) ⇒_{62 %, 21} Education(university)

The system is able to produce a sentence in natural language which expresses the meaning of the rule:

21 (i.e. 62 %) of the patients mainly sitting in their job and mainly sitting after the job also have reached university education.

The users have to only fill in linguistic data. The linguistic data are necessary for the transformation of the rule into natural language.

The system was tested for Czech, English and Finnish.

Files to download:

LM.4ft.AR2NL.zip	1.49 MB	January 30, 2005
The AR2NL system
templates.cs.zip	4.38 kB	December 8, 2004
Templates of XML data files for Czech.
templates.en.zip	3.00 kB	December 22, 2004
Templates of XML data files for English.
xxe2-config.zip	6.66 kB	December 8, 2004
A configuration file for XMLMind XML Editor.

History:

The 4ft-AR2NL module was designed by P. Strossa and implemented by Z. Černý in 2004.