Building Customized Search Engines

Abstract

Search engines are essential, ubiquitous tools for seeking information from the Internet. Prior research has also demonstrated that combining features of separate search engines often improves retrieval performance. However, such feature combination is often difficult, because developers don’t consider other developers when building their software. To facilitate the development of search engines, we propose a customized search engine approach to integrating appropriate components from multiple search engines. This article presents an interoperability architecture for building customized search engines. To achieve this, authors analyze existing search engines and decompose them into self-contained components that are classified into six categories. The Automated Software Development Environment for Information Retrieval (ASDEIR) was developed as the prototype of the proposed interoperability architecture to test its feasibility, robustness, and usefulness. ASDEIR incorporates intelligent features that detect and attempt to resolve conflicts between components.