​

module neuralforecast.utils

​

Global Variables

• AirPassengers
• unique_id
• ds
• y
• airline1_dummy
• airline2_dummy

​

function generate_series

generate_series(
    n_series: int,
    freq: str = 'D',
    min_length: int = 50,
    max_length: int = 500,
    n_temporal_features: int = 0,
    n_static_features: int = 0,
    equal_ends: bool = False,
    seed: int = 0
) → DataFrame

• n_series (int): Number of series for synthetic panel.
• freq (str, optional): Frequency of the data, panda’s available frequencies. Defaults to “D”.
• min_length (int, optional): Minimal length of synthetic panel’s series. Defaults to 50.
• max_length (int, optional): Maximal length of synthetic panel’s series. Defaults to 500.
• n_temporal_features (int, optional): Number of temporal exogenous variables for synthetic panel’s series. Defaults to 0.
• n_static_features (int, optional): Number of static exogenous variables for synthetic panel’s series. Defaults to 0.
• equal_ends (bool, optional): If True, series finish in the same date stamp ds. Defaults to False.
• seed (int, optional): Random seed for reproducibility. Defaults to 0.

• pd.DataFrame: Synthetic panel with columns [unique_id, ds, y] and exogenous.

​

function time_features_from_frequency_str

time_features_from_frequency_str(freq_str: str) → List[TimeFeature]

• freq_str (str): Frequency string of the form [multiple][granularity] such as “12H”, “5min”, “1D” etc.

• List[TimeFeature]: List of time features appropriate for the frequency.

​

function augment_calendar_df

augment_calendar_df(df, freq='H')

• Q - [month]
• M - [month]
• W - [Day of month, week of year]
• D - [Day of week, day of month, day of year]
• B - [Day of week, day of month, day of year]
• H - [Hour of day, day of week, day of month, day of year]
• T - [Minute of hour*, hour of day, day of week, day of month, day of year]
• S - [Second of minute, minute of hour, hour of day, day of week, day of month, day of year]

• df (pd.DataFrame): DataFrame to augment with calendar features.
• freq (str, optional): Frequency string for determining which features to add. Defaults to “H”.

• Tuple[pd.DataFrame, List[str]]: Tuple of (augmented DataFrame, list of feature column names).

​

function get_indexer_raise_missing

get_indexer_raise_missing(idx: Index, vals: List[str]) → List[int]

• idx (pd.Index): Index to search in.
• vals (List[str]): Values to find indices for.

• List[int]: List of index positions.

• ValueError: If any values are missing from the index.

​

function add_conformal_distribution_intervals

add_conformal_distribution_intervals(
    model_fcsts: <built-in function array>,
    cs_df: ~DFType,
    model: str,
    cs_n_windows: int,
    n_series: int,
    horizon: int,
    level: Optional[List[Union[int, float]]] = None,
    quantiles: Optional[List[float]] = None
) → Tuple[<built-in function array>, List[str]]

• model_fcsts (np.array): Model forecasts array.
• cs_df (DFType): DataFrame containing conformal scores.
• model (str): Model name.
• cs_n_windows (int): Number of conformal score windows.
• n_series (int): Number of series.
• horizon (int): Forecast horizon.
• level (Optional[List[Union[int, float]]], optional): Confidence levels for prediction intervals. Defaults to None.
• quantiles (Optional[List[float]], optional): Quantiles for prediction intervals. Defaults to None.

• Tuple[np.array, List[str]]: Tuple of (forecasts with intervals, column names).

​

function add_conformal_error_intervals

add_conformal_error_intervals(
    model_fcsts: <built-in function array>,
    cs_df: ~DFType,
    model: str,
    cs_n_windows: int,
    n_series: int,
    horizon: int,
    level: Optional[List[Union[int, float]]] = None,
    quantiles: Optional[List[float]] = None
) → Tuple[<built-in function array>, List[str]]

• model_fcsts (np.array): Model forecasts array.
• cs_df (DFType): DataFrame containing conformal scores.
• model (str): Model name.
• cs_n_windows (int): Number of conformal score windows.
• n_series (int): Number of series.
• horizon (int): Forecast horizon.
• level (Optional[List[Union[int, float]]], optional): Confidence levels for prediction intervals. Defaults to None.
• quantiles (Optional[List[float]], optional): Quantiles for prediction intervals. Defaults to None.

• Tuple[np.array, List[str]]: Tuple of (forecasts with intervals, column names).

​

function get_prediction_interval_method

get_prediction_interval_method(method: str)

• method (str): Name of the prediction interval method.

• Callable: The corresponding method function.

• ValueError: If the method is not supported.

​

function level_to_quantiles

level_to_quantiles(level: List[Union[int, float]]) → List[float]

• level (List[Union[int, float]]): List of confidence levels (e.g., [80, 90]).

• List[float]: List of corresponding quantiles.

​

function quantiles_to_level

quantiles_to_level(quantiles: List[float]) → List[Union[int, float]]

• quantiles (List[float]): List of quantiles (e.g., [0.1, 0.5, 0.9]).

• List[Union[int, float]]: List of corresponding confidence levels.

​

class TimeFeature

​

method __init__

__init__()

​

class SecondOfMinute

​

method __init__

__init__()

​

class MinuteOfHour

​

method __init__

__init__()

​

class HourOfDay

​

method __init__

__init__()

​

class DayOfWeek

​

method __init__

__init__()

​

class DayOfMonth

​

method __init__

__init__()

​

class DayOfYear

​

method __init__

__init__()

​

class MonthOfYear

​

method __init__

__init__()

​

class WeekOfYear

​

method __init__

__init__()

​

class PredictionIntervals

​

method __init__

__init__(n_windows: int = 2, method: str = 'conformal_distribution')

• n_windows (int, optional): Number of windows to evaluate. Defaults to 2.
• method (str, optional): One of the supported methods for the computation of prediction intervals: conformal_error or conformal_distribution. Defaults to “conformal_distribution”.