Mike Jackson

Ontology-based automatic classification for Web pages: design, implementation and evaluation

In recent years, we have witnessed continual growth in the use of ontologies in order to provide a mechanism to enable machine reasoning. This paper describes an automatic classifier, which focuses on the use of ontologies for classifying Web pages with respect to Dewey Decimal Classification (DDC) and Library of Congress Classification (LCC) schemes. Firstly, we explain how these ontologies can be built in a modular fashion, and mapped into DDC and LCC. Secondly, we propose the formal definition of a DDC-LCC and an ontology-classification-scheme mapping. Thirdly, we explain the way the classifier uses these ontologies to assist classification. Finally, an experiment in which the accuracy of the classifier was evaluated is presented. The experiment shows that our approach results an improved classification in terms of accuracy. This improvement, however, comes at a cost in a low coverage ratio due to incompleteness of the ontologies used.

Software measurement methods: recipes for success?

Abstract Measurement is a much advocated, yet infrequently applied technique of software engineering. A major contributory factor to this state of affairs is that the majority of software metrics are developed, collected and applied in a haphazard fashion. The result is metrics that frequently are poorly formulated, inappropriate to the specific needs and environment of the using organization and hard to analyse once collected. Over recent years a number of measurement methods—frameworks for developing and applying metrics—have been proposed and used to rectify this state of affairs. This paper describes and evaluates these various methods, identifies strengths and weaknesses leading to an agenda of further work. The requirements of a new measurement method are proposed and emphasizes increased application of measurement within context of software development processes.

A visual representation of search-engine queries and their results

The paper presents a visual technique for representing search engine results: the R-Wheel. The paper shows how the R-Wheel can be adapted to weight the relevance of individual terms of a multi-term query for input to a search engine. Five experiments are described requiring users to differentiate the importance of terms in a multi-term query; rank R-Wheels with the same and differing overall relevance; shade segments of an R-Wheel in proportion to the relevance of a query term; compare the R-Wheel with other representations and rank the representations in order of ease of interpretation. The paper concludes that the R-Wheel appears to be a valuable tool in visualising search engine results.

Thirty years (and more) of databases

Abstract This paper outlines the historical development of data management systems in order to identify the key issues for successful systems. It identifies the need for data independence and the embedding of structural and behavioural semantics in the database as key issues in the development of modern systems. Hierarchical, Network, Relational, Object-oriented and Object-relational data management systems are reviewed. A short summary of related research is given. The paper concludes with some speculation on the future directions that database technology might take.

A visual technique to indicate overall and component relevancy of search-engine results

The paper presents a tool for visualising search engine results: the R-Wheel. Four experiments are described. The first experiment looks at whether users can differentiate the importance of terms in a multi-term query. The second considers the use of the R-Wheel in recognising overall relevance. The third looks at the ability of users to use the R-Wheel to determine the most relevant results to a multi-term query. The fourth experiment compares the R-Wheel to traditional representations and asks users which representation they found the easiest to interpret. The paper concludes that users interpret the R-Wheel representation more correctly than others, and that users prefer the R-Wheel to other representations.