Optimizer-class: Class "Optimizer"
In frankongre/RcppGLM: Parallel Generalization Linear Mode

Class "Optimizer"

p: GLM object
w: coefficient matrix
w_prev: coefficient matrix of last iteration
g: gradient matrix
g_prev: gradient matrix of last iteration
f: objective function value
f_prev: objective function value of last iteration
d: direction coefficient matrix
direction: direction methods; character c("gradient", "newton", "stochastic gradient")
stepsize: stepsize methods; character c("wolfe", "backtrack")
maxiter: maximum iteration; default = 100
iter: iteration number
wtol: coefficient criteria for stop iteration
gtol: gradient tolerance
ftol: function value tolerance
t: stepsize