Description.md

Detail Parameter Settings / Default Setting

Decision Tree (DT)

• Maximum depth of tree

maxDepth = 5
opts = {'maxDepth':maxDepth, 'kfold':kfold}

Elastic Net (EN)

• EN contains several parameters

alpha  = 1.0      # constant that multiplies the penalty term
ratio  = 0.5      # Elastic Net mixing parameter
opts = {'alpha':alpha, 'ratio':ratio, 'kfold':kfold}

LASSO Regression

• Constant that multiplies the L1 term

alpha = 1
opts = {'alpha':alpha, 'kfold':kfold}

Neural Network (NN)

• NN contains several parameters

hls      = 100      # hidden layer size 
fun      = 'relu'   # activation function ( 'relu' or 'identity' or 'logistic' or 'tanh' )
max_iter = 100      # maximum iterations
opts = {'hls':hls, 'fun':fun, 'max_iter':max_iter, 'kfold':kfold}

RIDGE Regression

• Regularization strength

alpha = 1
opts = {'alpha':alpha, 'kfold':kfold}

Support Vector Regression (SVR)

• Kernel function

kernel = 'rbf'     # kernel ( 'linear' or 'poly' or 'rbf' or 'sigmoid' or 'precomputed' )
opts = {'kernel':kernel, 'kfold':kfold}