Abstract
With information retrieval systems, bridging the gap between the physical characteristics of data with the user perceptions is challenging. In order to address this challenge, employing user profiles to improve the retrieval accuracy becomes essential. However, the system performance may degrade due to inaccuracy of user profiles. Therefore, for an approach to be effective, it should offer a learning mechanism to correct user input errors. Focusing on an image retrieval application, we utilize the users’ relevance feedback to improve the profiles automatically using genetic algorithms (GA). Our experimental results indicated that the retrieval accuracy is significantly increased using the GA-based learning mechanism.