HTML has become widely used as a format for hypermedia documents. It specifies a model for text documents with hypertext links and non-text media objects. While HTML has served the needs of current hypertextual usage, it falls short of answering the anticipated demands of hypermedia environments. In particular, HTML enforces a single document that is closely tied to its presentation. As such, H...

The new wrapper model for extracting text data from HTML documents is introduced. In this model, an HTML file is considered as an ordered labeled tree. The learning algorithm takes the sequence of pairs of an HTML tree and a set of nodes The nodes indicate the labels to extract from the HTML tree. The goal of the learning algorithm is to output the wrapper which exactly extracts the labels from...

Search engines such as Google and MSN Search crawl and index files in Adobe’s Portable Document Format (PDF) alongside material in HTML. Google furthermore offers a View as HTML option for PDF that includes query term highlighting. The visual appearance of these HTML files converted from PDF is very poor. In this paper we claim that significant improvements to the quality of on-demand PDF to HT...

We present a new approach that automatically captures the semantic hierarchies in HTML tables, and semi-automatically integrates HTML tables belonging to a domain. It first automatically captures the attribute-value pairs in HTML tables by normalization and recognizing their headings. After generating global schema manually, it learns the lexical semantic sets and contexts, by which it then eli...

A standard feature in cataloging documents is the list of keywords. When the source documents are web pages, we can attempt to aid the cataloger by analyzing the page and presenting relevant support material. Since the keywords that occur in a document generally occur in keyphrases, and keyphrases provide contextual material for reviewing candidate keywords, they are a natural aggregate to extr...

Table is a commonly used presentation scheme, especially for describing relational information. Table understanding on the web has many potential applications including web mining, knowledge management, and web content summarization and delivery to narrow-bandwidth devices. Although in HTML documents tables are generally marked as elements, often the tag is used liberally to ach...

The number, the size, and the dynamics of Internet information sources bears abundant evidence of the need for automation in information extraction. This calls for representation formalisms that match the World Wide Web reality and for learning approaches and learnability results that apply to these formalisms. The concept of elementary formal systems is appropriately generalized to allow for t...