| Creator |
|
|
|
| Language |
|
| Date |
|
| Source Title |
|
| Vol |
|
| First Page |
|
| Last Page |
|
| Publication Type |
|
| Access Rights |
|
| Related DOI |
|
|
|
| Related URI |
|
|
|
| Relation |
|
|
|
| Abstract |
There are huge amount HTML pages on the Web. Many of them contains lists and tables. It is often the case that a line represents an instance of a record and that each column represents an attribute of... the record. Given query words, similar words can be obtained from a column if the query words are contained in the column. Given a record, as a pair of words, similar pairs can be obtained from other lines of the table. Therefore, the information of tables enables search of similar words and similar records. The granularity of these search is more fine compared to the conventional search engines whose result is a list of HTML pages. We are developing a search engine for tables on theWeb. It detects lists and tables by analyzing repeated structure in a HTML page. No examples are necessary in advance. No interaction is required in processing. No special knowledge or no natural language processing are necessary to detect the similarity of the data. This paper describes the structure of the system, the indexing mechanism of tables and typical applications of fine search using table index.show more
|
Export Mendeley
2003_e_1