GLM2¶

Supported HTTP methods and descriptions¶

GET GLM2

URL¶

http://<h2oHost>:<h2oApiPort>/GLM2.json

Input parameters¶

• offset, a Vec, <i>expert</i>

  Column to be used as an offset, if you have one.. Since version 1

• family, a Family, <i>critical</i>

  Family.. Since version 1

• link, a Link, <i>secondary</i>

  . Since version 1

• tweedie_variance_power, a double, <i>secondary</i>

  Tweedie variance power. Since version 1

• prior, a double, <i>expert</i>

  prior probability for y==1. To be used only for logistic regression iff the data has been sampled and the mean of response does not reflect reality.. Since version 1

• disable_line_search, a boolean, <i>expert</i>

  disable line search in all cases.. Since version 1

• n_folds, a int, <i>critical</i>

  validation folds. Since version 1

• alpha, a double[], <i>secondary</i>

  distribution of regularization between L1 and L2.. Since version 1

• lambda, a double[], <i>secondary</i>

  regularization strength. Since version 1

• lambda_search, a boolean, <i>secondary</i>

  use lambda search starting at lambda max, given lambda is then interpreted as lambda min. Since version 1

• nlambdas, a int, <i>expert</i>

  number of lambdas to be used in a search. Since version 1

• lambda_min_ratio, a double, <i>expert</i>

  min lambda used in lambda search, specified as a ratio of lambda_max. Since version 1

• max_predictors, a int, <i>expert</i>

  lambda_search stop condition: stop training when model has more than than this number of predictors (or don’t use this option if -1).. Since version 1

• strong_rules, a boolean, <i>secondary</i>

  use strong rules to filter out inactive columns. Since version 1

• standardize, a boolean, <i>critical</i>

  Standardize numeric columns to have zero mean and unit variance.. Since version 1

• intercept, a boolean, <i>critical</i>

  Include intercept term in the model.. Since version 1

• non_negative, a boolean, <i>critical</i>

  Restrict coefficients to be non-negative.. Since version 1

• beta_constraints, a Frame, <i>expert</i>

  lower bounds for coefficients. Since version 1

• use_all_factor_levels, a boolean, <i>secondary</i>

  By default, first factor level is skipped from the possible set of predictors. Set this flag if you want use all of the levels. Needs sufficient regularization to solve!. Since version 1

• variable_importances, a boolean, <i>secondary</i>

  Compute variable importances for input features. NOTE: If use_all_factor_levels is off the importance of the base level will NOT be shown.. Since version 1

• beta_epsilon, a double, <i>secondary</i>

  beta_eps. Since version 1

• max_iter, a int, <i>critical</i>

  max-iterations. Since version 1

• higher_accuracy, a boolean, <i>secondary</i>

  use line search (slower speed, to be used if glm does not converge otherwise). Since version 1

Output JSON elements¶

• _wgiven, a double[]

  . Since version 1, secondary

• _proximalPenalty, a double

  . Since version 1, secondary

• _beta, a double[]

  . Since version 1, secondary

• _runAllLambdas, a boolean

  . Since version 1, secondary

• tweedie_link_power, a double

  Tweedie link power. Since version 1, secondary

• lambda_max, a double

  lambda_value max. Since version 1, secondary

HTTP response codes¶

200 OK Success and error responses are identical.