2013 Fiscal Year Final Research Report
Computer based methods for automatic sampling and correspondence table generation for similar regulations among multiple municipalities
| Project/Area Number |
21500253
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Library and information science/Humanistic social informatics
|
|---|
| Research Institution | Dokkyo University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TAKENAKA Yoishi 大阪大学, 大学院情報科学研究科, 准教授 (00324830)
|
|---|
| Project Period (FY) |
2009-04-01 – 2014-03-31
|
| Keywords | 言語処理 / 都道府県 / 道州制 / 条例 / 計算機 / 自動抽出 / アライメント / 例規 |
|---|
| Research Abstract |
For the purpose of clarifying differences in regulations among municipalities, we studied methods to automatically sample similar rules by computer and to automatically generate the clause-correspondence table between two municipalities. We initially converted codes of all Japanese prefectural governments, which are open on Website, into XML documents and demonstrated that TF-IDF method is effective to automatically generate the clause-correspondence table. However, the TF-IDF method requires too long. We investigated alternative methods with shorter computational time and identified that the length of longest common subsequence (LCS) among clause-headings is the most effective with the shortest time. We also clarified that the length of LCS is effective to automatically extract similar regulations from other prefectural governments. Thus, we concluded that our study should contribute to reducing workloads of legal sections of municipalities, especially when they would merge.
|
