| AITEC Contract Research Projects in FY1996 : Software |
Machine : UNIX machine Environment : UNIX (SunOS Release 5.4) Language : SICStus-Prolog 3 #3
The software performs legal analogies which are dependent on its goal. Given a goal that is related to some ground of legal rule, the system detects similarities among legal concepts which share the same goal and proof. The computed similarities between the concepts are represented by groups of concepts. Each group shows a similarity class of concepts sharing the same ground and proof.
Application domain: Analogy in legal reasoning using order-sorted logic.
Functionality: Detecting similarities among sorts according to a goal.
Input: Order-sorted representation of a legal domain in the form of
Horn-clauses, and a goal to be proved under the domain knowledge.
Main Usage: Analogical reasoning in legal reasoning, argumentation,
and debugging of incomplete knowledge.
Readme-E ........... List of files (in English)
Readme-J ........... List of files (in Japanese)
INSTALL ............ Install manual (in Japanese)
manual.txt ......... User's manual (in Japanese)
name.txt ........... Copyright (in Japanese)
pub.txt ............ List of the publications (in Japanese)
spec.txt ........... Specifications of this system (in Japanese)
src/
newgda.pl ........ Source program (main module)
user.pl .......... Source program (interface module)
sample/
demo.kb .......... Sample knowledge
Source Code : 71 KB (2405 step)
Documents : Users manual (text file in Japanese)
* This software detects similarities according to the ground of
legal rules which represent the context-dependency of reasoning.
* This software detects all relevant similarities which retain
proof of the ground of legal rules.
* This software uses several constraints to considerably reduce
the huge space of possible similarities.
These originate from a knowledge of the terminology of legal concepts.
* This software also contains a simple legal reasoning system.
The system can perform legal reasoning through direct analogies
of detected similarities.
www-admin@icot.or.jp