A framework for specifying explicit bias for revision of approximate information extraction rules

Abstract

Information extraction is one of the most important techniques used in Text Mining. One of the main problems in building information extraction (IE) systems is that the knowledge elicited from domain experts tends to be only approximately correct. In addition, the knowledge acquisition phase for building IE rules usually takes a tremendous amount of time on the part of the expert and of the linguist creating the rules. We therefore need an effective means of revising our IE rules whenever we discover such an inaccuracy. The IE revision problem is how best to go about revising a deficient IE rules using information contained in examples that expose inaccuracies. The revision process is very sensitive to implicit and explicit biases encoded in the specific revision algorithm employed. In a sense, each revision algorithm must provide two forms of biases: bias as to the place of the revision and bias as to the type of the revision that should be performed. In this paper we present a framework for writing approximate IE rules that are provided with explicit bias. The proposed framework can be used by many existing revision algorithms. The purpose of the revision bias framework is to allow the user to declare his own bias in a simple and structured way, i.e. to express the conditions placed on the domain knowledge for a given revision operator to be applied. This language extends and generalizes the work reported in [Feldman et. al. 1993]. It attacks the problem of writing IE rules from a novel perspective, one which enables a much faster development of IE systems.

Mining the Biomedical Literature in the Genomic Era: An Overview

“…Based on actual empirical evaluation, it was found that it is enough to focus just on the core constituents of sentences and use shallow parsing augmented by smart skips. These skips enable the information extraction engine to skip irrelevant parts, and focus just on the important phrases of each sentence (Appelt and Israel, 1999;Feldman et al, 2001;Feldman et al, 2000). Researchers have attempted before to use full parsing as a component in their information systems and have concluded that it was not worthwhile to invest the extra effort.…”

Discovering unexpected information from your competitors' web sites

“…In keyword-based search, the user specifies some keywords, and a search engine (e.g., Yahoo, Excite, Alta Vista, and Google) finds those Web pages that contain the keywords, and ranks them according to various measures. In Web information extraction [e.g., 2,13,9,15,8], a wrapper or a specific extraction procedure is built automatically or manually for a Web page to extract some specific pieces of information requested by the user, e.g., extracting the prices of some products. User preference based approaches are commonly used in push type of systems [e.g., 26], where the user specifies what categories of information are interesting to him/her.…”

“…In Web query based approaches, database query language such as SQL is extended and modified so that it can be used to query semi-structured information resources, XML documents and Web pages [e.g., 19,12,6,8]. Web resource discovery aims to find resources (Web pages) related to the user requests [e.g., 16,7,8,9,10,13]. This approach uses techniques such as link analysis, and text So far, we have not given extra considerations to metadata [11,22] and hyperlinks ll6], which will be studied in our future work.…”

“…Wrapper-based approaches [e.g., 2,13,9,15] enable the user to extract specific pieces of information from targeted Web pages. Web query languages [e.g., 19,12,6] give the user the opportunity to query the Web.…”

Intelligent Web Information Mapping with HyperMap

“…Web query languages allow the user to query Web pages ( [3,8]). Wrapper-based information extraction enables the user to extract specific pieces of information from targeted Web sites ( [6,7]). The user preference approaches provide the information to the user according to users interests and requirements [13].…”