AMK

Additive Multiplicative Kernel Regression (AMK) based power curve estimation.

from dwse import AMK
model = AMK(X_train, y_train)
prediction = model.predictions
class dswe.AMK.AMK(X_train, y_train, X_test, bw='dpi', n_multi_cov=3, fixed_cov=[0, 1], cir_cov=None)[source]
Parameters

  • X_train (np.ndarray or pd.DataFrame) – A matrix or dataframe of input variable values in the training dataset.

  • y_train (np.array) – A numeric array for response values in the training dataset.

  • X_test (np.ndarray or pd.DataFrame) – A matrix or dataframe of test input variable values to compute predictions.

  • bw (string or int) – A numeric array or a character input for bandwidth. If character, bandwidth computed internally; the input should be either ‘dpi’ or ‘dpi_gap’. Default value is ‘dpi_gap’.

  • n_multi_cov (int) – An integer or a character input specifying the number of multiplicative covariates in each additive term. Default is 3 (same as Lee et al., 2015). The character inputs can be: ‘all’ for a completely multiplicative model, or ‘none’ for a completely additive model. Ignored if the number of covariates is 1.

  • fixed_cov (list) – An integer list specifying the fixed covariates column number(s). Ignored if n_multi_cov is set to ‘all’ or ‘none’ or if the number of covariates is less than 3. Default value is [0,1].

  • cir_cov (list or int) – A list specifying the circular covariates column number(s) in X_train, An integer when only one circular covariates present. Default value is None.

Returns

self with trained parameter values.

  • predictions: stored numeric array of model output at the data points in X_test.

Return type

AMK

Reference

Lee, Ding, Genton, and Xie, 2015, “Power curve estimation with multivariate environmental factors for inland and offshore wind farms,” Journal of the American Statistical Association, Vol. 110, pp. 56-67.