tigerforecast.utils.autotuning.grid_search.GridSearch

class tigerforecast.utils.autotuning.grid_search.GridSearch

Description: Implements the equivalent of an AR(p) method - predicts a linear combination of the previous p observed values in a time-series

__init__()

Initialize self. See help(type(self)) for accurate signature.

Methods

__init__()

Initialize self.

search(method_id, method_params, problem_id, …)

Description: Search for optimal method parameters :param method_id: id of method :type method_id: string :param method_params: initial method parameters dict (updated by search space) :type method_params: dict :param problem_id: id of problem to try on :type problem_id: string :param problem_params: problem parameters dict :type problem_params: dict :param loss: a function mapping y_pred, y_true -> scalar loss :type loss: function :param search_space: dict mapping parameter names to a finite set of options :type search_space: dict :param trials: number of random trials to sample from search space / try all parameters :type trials: int, None :param smoothing: loss computed over smoothing number of steps to decrease variance :type smoothing: int :param min_steps: minimum number of steps that the method gets to run for :type min_steps: int :param verbose: if 1, print progress and current parameters :type verbose: int