Abstract
This paper is a presentation of an approach to the prediction of central vestibular diseases or peripheral vestibular diseases, that are Meniere's disease, sudden deafness, vestibular neuronitis, benign paroxysmal positional vertigo, and drug ototoxicity, from mathematical statistical point of view. An analysis was performed by the quantification method for multi-dimensional qualitative data based on 13 otoneurologic tests and sex. We got 5 axes (roots) because there are 6 groups of outside criterion. We chose the first, the second, and the third axes in discrimination study because of their large value of correlation ratio. We could estimate that the first axis divided the group of peripheral vestibular diseases, and the group of central vestibular diseases, and that the OKP test was the most significant test in discriminating between them. Second axis divided the group of Meniere's disease, sudden deafness, and drug ototoxicity, and the group of vestibular neuronitis, and benign paroxysmal positional vertigo. The hearing test was the most significant test on the second axis. The third axis divided the vestibular neuronitis, and benign paroxysmal positional vertigo, and the caloric test was the most significant in discrimination.
