紀要論文

語句関連性とベイズ推論による対話型法律相談支援システム

Words Correlation and Bayesian Inference based　Interactive Legislation Consultation Support System

jpn

18810667

https://doi.org/10.15002/00009879

法政大学大学院情報科学研究科

8

207

210

2013-03

Version of Record

With increasing information and continuously updating information, it becomes difficult to find appropriate information needed. Traditionally, people consults with a lawyer when they are involved in a case related to a legal issue but it takes time and cost rather money. This paper describes a legal issue consultant support system which provides three functions: making graph expressed in the word relations; learning to know what a user wants to via the interactive process; and suggesting the related contents derived from Bayesian inference. Through interactions with the user, the system collects a sets of keywords related user needs, the system tries to find out what the user wants and retrieves them from the related graph database using Bayesian inference. Even for a beginner, it is easy to use since its operations are simple without requiring technical words. Moreover, the system is easy for administrator to manage. On the whole, the system is aimed at learning what a user needs, providing helpful information, and helping them to understand legislation.

Article

AA12222297