Figure 2
From: Classification of heterogeneous microarray data by maximum entropy kernel
Schematic view of the entire process of microarray classification in the ME kernel algorithm. The input vectorial data are first converted into distance matrix to provide constraints D ij . Then, entropy of a kernel matrix is maximized under the constraints, generating an optimal kernel matrix that is guaranteed to be positive semidefinite. Then, the SVM learns the classification boundary from the kernel matrix and classifies test samples.