V3 N1 Paper 3
Annals of the MS in Computer Science and Information Systems at UNC Wilmington
Spring 2009

Evaluation of Spelling Correction and Concept-Based Searching Models  

Royce Nobles

Committee

Curry Guinn (chair)
Tom Janicki
Darwin Dennison

Abstract

Much advancement has taken place in the field of Information Retrieval toward improving the efficiency and ease with which documents can be located. While research has improved the performance of text searches on massive, domain independent databases (e.g., Google™), most domain-specific applications available to consumers use more limited search techniques. In particular, nutrition software applications for logging and analyzing an individual’s daily food consumption commonly employ simple Boolean text matching to locate food items from within a database, often resulting in long search times and difficulty locating foods. DINE Healthy is a nutrition assessment and diet improvement software program developed by DINE Systems, Inc. and widely used in college and university nutrition classes. The current version (7.0) accomplishes food searching by matching user query strings with food item descriptions using the common Boolean Search approach. This research evaluates the effectiveness of enhancing the DINE Healthy food search system with advanced Information Retrieval techniques including spelling error correction and concept-based searching. To facilitate independent evaluation of these techniques, two additional food search versions are developed for use with the DINE Healthy software. The first extends the current system with only spelling error correction while the second includes both spelling error correction and concept-based searching. Analysis of usage data from a population of university students using each food search version demonstrates that improvements in speed of data entry, a reduction in the rate of failed searches, and a decreased number of queries required to locate food items can be achieved using the enhanced search versions.

download (pdf)

Recommended Citation: Nobles, R., Guinn, C, Janicki, T., Dennison, D. (2009) Evaluation of Spelling Correction and Concept-Based Searching Models. Annals of the Master of Science in Computer Science and Information Systems at UNC Wilmington, 3(1) paper 3. http://csbapp.uncw.edu/data/mscsis/full.aspx.

V3 N1 Paper 3
Annals of the MS in Computer Science and Information Systems at UNC Wilmington
Spring 2009