Comparison of linear and non-linear classifiers for oral cancer screening by optical spectroscopy.

We compared the use of linear discriminant analysis and an artificial neural network for classifying optical spectral measurements of oral sites. Consistent with studies of optical spectroscopy for diagnosis of other forms of cancer, our results suggest that a non-linear classifier may be warranted. Our study also demonstrated that the classifiers were better able to distinguish between sites that were more histopathologically distinct.