Local linear logistic eliminates bias at boundries of domain. It uses weighted least squares, determining weights from the kernel. Args: x (np.array): Training data of shape[1, n_features] note: currently only single samples can be predi
(self, x, kernel, gamma, reg_param=0,
iterations=10, alpha=0.001)
source not stored for this graph (policy: none)
nothing calls this directly
no test coverage detected