Mining tinnitus data based on clustering and new temporal features

Abstract

Tinnitus problems affect a significant portion of the population and are difficult to treat. Sound therapy for Tinnitus is a promising, expensive, and complex treatment, where the complete process may span from several months to a couple of years. The goal of this research is to explore different combinations of important factors leading to a significant recovery, and their relationships to different category of Tinnitus problems. Our findings are extracted from the data stored in a clinical database, where confidential information had been stripped off. The domain knowledge spans different disciplines such as otology as well as audiology. Complexities were encountered with temporal data and text data of certain features. New temporal features together with rule generating techniques and clustering methods are presented with a ultimate goal to explore the relationships among the treatment factors and to learn the essence of Tinnitus problems. © 2011 Springer-Verlag Berlin Heidelberg.