An analog system-on-chip for kernel-based pattern classification and sequence estimation is presented. State transition probabilities conditioned on input data are generated by an...
We proposed a new approach to compare profiles when the correlations among attributes can be represented as a tree. To account for these correlations, the profile is extended with...
Infection by the human papillomavirus (HPV) is associated with the development of cervical cancer. HPV can be classified to highand low-risk type according to its malignant potenti...
The choice of the kernel function which determines the mapping between the input space and the feature space is of crucial importance to kernel methods. The past few years have se...
In this paper, a novel method for reducing the runtime complexity of a support vector machine classifier is presented. The new training algorithm is fast and simple. This is achiev...