Information retrieval with concept ontology for domain-specific text

In an age of information boom, efficient retrieval of information is becoming more important. Enormous amount of information can sometimes cause over-load for people. Moreover, traditional search engines only return users with a ranked list of documents without a grand overview of information. To address users' growing information needs, we proposed an information retrieval solution with the use of concept ontology. By integrating information retrieval with ontology, users can effectively navigate among different documents and have a quick grasp of the information contained in the documents. A proof-of-concept web application, named DSPLearn, was developed in the domain of digital signal processing. It integrates traditional keyword search with the idea of concept ontology. The technologies behind DSPLearn are generic and can be applied to any kind of text and any other knowledge bases. DSPLearn supports efficient search of PDF documents. It generates a concept tree based on the search results for a query, from which users can filter the results. It also allows highlighting of terms that are mapped to some user-selected concepts in a PDF document. An n-th match approach was proposed to locate an exact term in a document. With rapid information growth, the idea of concept ontology is promising. Ontology will play a significant part in building a Semantic Web - a Web of data.Bachelor of Engineering (Computer Science