An Improved Kernel Credal Classification Algorithm Based on Regularized Mahalanobis Distance: Application to Microarray Data Analysis

Algorithm 1

Mahalanobis kernel credal classification rule (MKCCR).

Input

X: Dataset

c: Clusters Number

β: The Tuning Parameter

γ: The Tuning Parameter for objects in Meta-cluster

h: Hyper Parameter of the Gaussian Kernel

: Regularization Parameter (τ=0.01)

Initialization

Choose initial cluster centers from the objects

Calculate the distance between data and centers using equation (23)

Construct the bba’s using equations (13), (14) and (16)