predict_trend¶
-
predict_trend
(fit_result: Dict[str, Union[lmfit.model.ModelResult, pandas.core.frame.DataFrame]], days_to_predict: int = 30, func_options: dict = {}, param_inverted_stderr: Iterable[str] = [], brute_force_extrema: bool = False) → pandas.core.frame.DataFrame[source]¶ Generic function to predict a trend from fitted data
- Parameters
fit_result (Dict[str, Union[lmfit.model.ModelResult, pd.DataFrame]]) – result of fit_data_model or its implementation
days_to_predict (int, optional) – number of days to predict a trend for, by default 30
func_options (dict, optional) – options for the function of model, by default {}
param_inverted_stderr (Iterable[str], optional) – iterable of parameternames with should be inverted, to calculate the extrema. , by default []
brute_force_extrema (bool, optional) – Whether or not to calculate supremum and infimum from all permutations of adding and subtracting the errors from the parameters. For some functions, i.e. the logistic curve, this is needed, since simply adding or subtracting the errors from the parameter can lead to supremum and/or infimum to cross the result with the exact parameters., by default False
- Returns
DataFrame with columns “date”, “trend”, “trend_sup” and “trend_inf”
- date: pd.Datetime
date of the values
- trend: float
predicted trend
- trend_sup: float
supremum of the trend
- trend_inf: float
infimum of the trend
- Return type
pd.DataFrame
See also
fit_data_model()
,calc_extrema()
,params_to_df()