Table 2 Workflow of DNN classifier

Input: Gene data matrix A ∈ {0, 1}^m × n after CGF and ISR, where rows and columns correspond to samples and genes, respectively; max training epoch E _max.

1: Training: for each training epoch e ≤ E _max:

 (a) For each sample a _i  = A(i, :):

  i. Conduct feed-forwarding and compute the loss J;

  ii. Conduct back-propagation to update the W and b

2: Testing: for each sample a _i  = A(i, :):

 (a) Conduct feed-forwarding and get softmax probability P;

 (b) Adopt the cancer type correspond to max(P) as the result of a _i .

Output: Trained network model (training) or classification results for the samples (testing).