DMCA
Extracting structured data from web pages (2003)
Cached
Download Links
- [csdl.computer.org]
- [www.dbai.tuwien.ac.at]
- [research.microsoft.com]
- [research.microsoft.com]
- DBLP
Other Repositories/Bibliography
| Venue: | In ACM SIGMOD |
| Citations: | 310 - 0 self |
BibTeX
@INPROCEEDINGS{Arasu03extractingstructured,
author = {Arvind Arasu},
title = {Extracting structured data from web pages},
booktitle = {In ACM SIGMOD},
year = {2003},
pages = {337--348}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Many web sites contain a large collection of “structured” web pages. These pages encode data from an underlying structured source, and are typically generated dynamically. An example of such a collection is the set of book pages in Amazon. There are two important characteristics of such a collection: first, all the pages in the collection contain structured data conforming to a common schema; second, the pages are generated using a common template. Our goal is to automatically extract structured data from a collection of pages described above, without any human input like manually generated rules or training sets. Extracting structured data gives us greater querying power over the data and is useful in information integration systems. Most of the existing work on extracting structured data assumes significant human input, for example, in form of training examples of the data to be extracted. To the best of our knowledge, ROADRUNNER project is the only other work that tries to automatically extract structured data. However, ROADRUNNER makes several simplifying assumptions. These assumptions and their implications are discussed in our paper [2]. Structured data denotes data conforming to a schema or type. We borrow the definition of complex types from [1]. Any value conforming to a schema is an instance of the schema. For example, the schema ¡ £ ¥ § © ¥ � § ¥ � represents a tuple of � attributes. The first and third attributes are “atomic”; the second attribute is a set of atomic values. The value denotes an instance of schema. A template is a pattern that describes how instances of a schema are encoded. An example template for schema above � is where each letter denotes a string. Template � encodes the first attribute of between strings � and �, the second between �
Keyphrases
web page book page information integration system large collection example template collection contain structured web page common schema complex type data assumes significant human input second attribute roadrunner project important characteristic structured data denotes data third attribute many web site human input atomic value first attribute common template
