Projects
← Back to all projects2025 / Archived
Query Errors and Cognitive Variance.
Short paper evaluating how dyslexia and Parkinson induced errors in user queries affect precision and recall of classical IR systems, and how a simple Levenshtein distance normalisation step recovers most of the lost retrieval quality.
- Role
- Author
- Status
- Archived
- Year
- 2025
- Links
A short study at the intersection of information retrieval and cognitive variation. The work models systematic input errors typical of dyslexia and the motor effects of Parkinson, and measures their impact on the precision and recall of a conventional retrieval system.
The result: a simple normalisation layer based on Levenshtein distance recovers most of the original retrieval quality without altering the retrieval model itself. Methodologically the paper is an evaluation study, not a new model.
The work signals my research interest in information retrieval and NLP beyond product engineering.