Abstract

In this study a supervised classifier based on the kernel implementation of the Bayes rule is introduced. The proposed technique first suggests an implicit nonlinear transformation of the data into a feature space and then seeks to fit normal distributions having a common covariance matrix onto the mapped data. The use of kernel concept in this process gives us the flexibility required to model complex data structures that originate from a wide-range of class conditional distributions. Although the decision boundaries in the new feature space are piece-wise linear, these corresponds to powerful nonlinear boundaries in the original input space. For the data we considered we have obtained some encouraging results.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call