A Framework of a interactive consultation system for suit using newspaper articles
| Project/Area Number |
19500087
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Media informatics/Database
|
|---|
| Research Institution | Iwate Prefectural University |
|---|
Principal Investigator |
KUREMATSU Masaki Iwate Prefectural University, ソフトウェア情報学部, 准教授 (00305286)
|
|---|
| Project Period (FY) |
2007 – 2009
|
| Project Status |
Completed (Fiscal Year 2009)
|
| Budget Amount *help |
¥2,860,000 (Direct Cost: ¥2,200,000、Indirect Cost: ¥660,000)
Fiscal Year 2009: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2008: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2007: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
|
|---|
| Keywords | 情報検索 / 法情報 / 情報抽出 / 文書解析 / 法情報処理 / 照応解析 / 新聞記事解析 |
|---|
| Research Abstract |
I proposed the following algorithm for a lawsuit consultation support system using newspaper articles. This algorithm has 3 Steps. In step.1, it tries to decode clipped words using reference terms. In order to do so, it focuses phrases which have same words in a clipped word and appear before the clipped word. It replaces the clipped word by the nearest phrase. In step.2, it divides an article into title, lead, first paragraph and rest of the paragraph. Then, it extracts words from each block. In step.3, it retrieves articles from previous articles based on words of each block. When it retrieves articles, it calculates the coincidence rate and the share rate of words in each block. If both rates of an article over these thresholds, it returns the article and adds it to a collection of articles about same topics. When a user asks this system, it extracts some sentences which include words in a question. I evaluated the performance of this algorithm about traffic accident cases on with desk check. Consequently, I have to improve the retrieval performance of this algorithm.
|
Report
(4 results)
Research Products
(5 results)
