CSVR

class pystoned.CSVR.CSVR(y, x, fun='prod', epsilon=0.01, C=2)[source]

Convex Support Vector Regression (CSVR)

__init__(y, x, fun='prod', epsilon=0.01, C=2)[source]

CSVR model

Parameters:

  • y (float) – output variable.

  • x (float) – input variables.

  • epsilon (float) – epsilon-loss function.

  • C (float) – Regularization parameter.

  • fun (String, optional) – FUN_PROD (production frontier) or FUN_COST (cost frontier). Defaults to FUN_PROD.

display_alpha()[source]

Display alpha value

display_beta()[source]

Display beta value

display_status()[source]

Display the status of problem

get_alpha()[source]

Return alpha value by array

get_beta()[source]

Return beta value by array

get_predict(x_test)[source]

Return the estimated function in testing sample

get_status()[source]

Return status

optimize(email='local', solver=None)[source]

Optimize the function by requested method

Parameters:

  • email (string) – The email address for remote optimization. It will optimize locally if OPT_LOCAL is given.

  • solver (string) – The solver chosen for optimization. It will optimize with default solver if OPT_DEFAULT is given.