autots package

Module contents

Automated Time Series Model Selection for Python

https://github.com/winedarksea/AutoTS

autots.load_daily()

4 series of sample daily data from late 2019.

autots.load_monthly()

Federal Reserve of St. Louis monthly economic indicators.

autots.load_yearly()

Federal Reserve of St. Louis annual economic indicators.

autots.load_hourly()

Traffic data from the MN DOT via the UCI data repository.

autots.load_weekly()

Weekly petroleum industry data from the EIA.

class autots.AutoTS(forecast_length: int = 14, frequency: str = 'infer', prediction_interval: float = 0.9, max_generations: int = 5, no_negatives: bool = False, constraint: float = None, ensemble: str = 'auto', initial_template: str = 'General+Random', random_seed: int = 2020, holiday_country: str = 'US', subset: int = None, aggfunc: str = 'first', na_tolerance: float = 1, metric_weighting: dict = {'containment_weighting': 0, 'contour_weighting': 0, 'mae_weighting': 2, 'rmse_weighting': 2, 'runtime_weighting': 0, 'smape_weighting': 10, 'spl_weighting': 1}, drop_most_recent: int = 0, drop_data_older_than_periods: int = 100000, model_list: str = 'default', num_validations: int = 2, models_to_validate: float = 0.15, max_per_model_class: int = None, validation_method: str = 'even', min_allowed_train_percent: float = 0.5, remove_leading_zeroes: bool = False, model_interrupt: bool = False, verbose: int = 1, n_jobs: int = None)

Bases: object

Automate time series modeling using a genetic algorithm.

Parameters
  • forecast_length (int) – number of periods over which to evaluate forecast. Can be overriden later in .predict().

  • frequency (str) – ‘infer’ or a specific pandas datetime offset. Can be used to force rollup of data (ie daily input, but frequency ‘M’ will rollup to monthly).

  • prediction_interval (float) – 0-1, uncertainty range for upper and lower forecasts. Adjust range, but rarely matches actual containment.

  • max_generations (int) – number of genetic algorithms generations to run. More runs = longer runtime, generally better accuracy.

  • no_negatives (bool) – if True, all negative predictions are rounded up to 0.

  • constraint (float) – when not None, use this value * data st dev above max or below min for constraining forecast values. Applied to point forecast only, not upper/lower forecasts.

  • ensemble (str) – None, ‘simple’, ‘distance’

  • initial_template (str) – ‘Random’ - randomly generates starting template, ‘General’ uses template included in package, ‘General+Random’ - both of previous. Also can be overriden with self.import_template()

  • random_seed (int) – random seed allows (slightly) more consistent results.

  • holiday_country (str) – passed through to Holidays package for some models.

  • subset (int) – maximum number of series to evaluate at once. Useful to speed evaluation when many series are input.

  • aggfunc (str) – if data is to be rolled up to a higher frequency (daily -> monthly) or duplicate timestamps are included. Default ‘first’ removes duplicates, for rollup try ‘mean’ or np.sum. Beware numeric aggregations like ‘mean’ will not work with non-numeric inputs.

  • na_tolerance (float) – 0 to 1. Series are dropped if they have more than this percent NaN. 0.95 here would allow series containing up to 95% NaN values.

  • metric_weighting (dict) – weights to assign to metrics, effecting how the ranking score is generated.

  • drop_most_recent (int) – option to drop n most recent data points. Useful, say, for monthly sales data where the current (unfinished) month is included.

  • drop_data_older_than_periods (int) – take only the n most recent timestamps

  • model_list (list) – list of names of model objects to use

  • num_validations (int) – number of cross validations to perform. 0 for just train/test on final split.

  • models_to_validate (int) – top n models to pass through to cross validation. Or float in 0 to 1 as % of tried. 0.99 is forced to 100% validation. 1 evaluates just 1 model. If horizontal or probabilistic ensemble, then additional min per_series models above the number here may be added to validation.

  • max_per_model_class (int) – of the models_to_validate what is the maximum to pass from any one model class/family.

  • validation_method (str) – ‘even’, ‘backwards’, or ‘seasonal n’ where n is an integer of seasonal ‘backwards’ is better for recency and for shorter training sets ‘even’ splits the data into equally-sized slices best for more consistent data ‘seasonal n’ for example ‘seasonal 364’ would test all data on each previous year of the forecast_length that would immediately follow the training data.

  • min_allowed_train_percent (float) – percent of forecast length to allow as min training, else raises error. 0.5 with a forecast length of 10 would mean 5 training points are mandated, for a total of 15 points. Useful in (unrecommended) cases where forecast_length > training length.

  • remove_leading_zeroes (bool) – replace leading zeroes with NaN. Useful in data where initial zeroes mean data collection hasn’t started yet.

  • model_interrupt (bool) – if False, KeyboardInterrupts quit entire program. if True, KeyboardInterrupts attempt to only quit current model. if True, recommend use in conjunction with verbose > 0 and result_file in the event of accidental complete termination.

  • verbose (int) – setting to 0 or lower should reduce most output. Higher numbers give more output.

  • n_jobs (int) – Number of cores available to pass to parallel processing. A joblib context manager can be used instead (pass None in this case).

best_model

DataFrame containing template for the best ranked model

Type

pandas.DataFrame

regression_check

If True, the best_model uses an input ‘User’ future_regressor

Type

bool

export_template(filename, models: str = 'best', n: int = 5, max_per_model_class: int = None, include_results: bool = False)

Export top results as a reusable template.

Parameters
  • filename (str) – ‘csv’ or ‘json’ (in filename). None to return a dataframe and not write a file.

  • models (str) – ‘best’ or ‘all’

  • n (int) – if models = ‘best’, how many n-best to export

  • max_per_model_class (int) – if models = ‘best’, the max number of each model class to include in template

  • include_results (bool) – whether to include performance metrics

fit(df, date_col: str = None, value_col: str = None, id_col: str = None, future_regressor=[], weights: dict = {}, result_file: str = None, grouping_ids=None)

Train algorithm given data supplied.

Parameters
  • df (pandas.DataFrame) – Datetime Indexed dataframe of series, or dataframe of three columns as below.

  • date_col (str) – name of datetime column

  • value_col (str) – name of column containing the data of series.

  • id_col (str) – name of column identifying different series.

  • future_regressor (numpy.Array) – single external regressor matching train.index

  • weights (dict) – {‘colname1’: 2, ‘colname2’: 5} - increase importance of a series in metric evaluation. Any left blank assumed to have weight of 1.

  • result_file (str) – results saved on each new generation. Does not include validation rounds. “.csv” save model results table. “.pickle” saves full object, including ensemble information.

  • grouping_ids (dict) – currently a one-level dict containing series_id:group_id mapping.

import_results(filename)

Add results from another run on the same data.

Input can be filename with .csv or .pickle. or can be a DataFrame of model results or a full TemplateEvalObject

import_template(filename: str, method: str = 'Add On', enforce_model_list: bool = True)

Import a previously exported template of model parameters. Must be done before the AutoTS object is .fit().

Parameters
  • filename (str) – file location (or a pd.DataFrame already loaded)

  • method (str) – ‘Add On’ or ‘Only’

  • enforce_model_list (bool) – if True, remove model types not in model_list

predict(forecast_length: int = 'self', prediction_interval: float = 'self', future_regressor=[], hierarchy=None, just_point_forecast: bool = False, verbose: int = 'self')

Generate forecast data immediately following dates of index supplied to .fit().

Parameters
  • forecast_length (int) – Number of periods of data to forecast ahead

  • prediction_interval (float) – interval of upper/lower forecasts. defaults to ‘self’ ie the interval specified in __init__() if prediction_interval is a list, then returns a dict of forecast objects.

  • future_regressor (numpy.Array) – additional regressor, not used

  • hierarchy – Not yet implemented

  • just_point_forecast (bool) – If True, return a pandas.DataFrame of just point forecasts

Returns

Either a PredictionObject of forecasts and metadata, or if just_point_forecast == True, a dataframe of point forecasts

results(result_set: str = 'initial')

Convenience function to return tested models table.

Parameters

result_set (str) – ‘validation’ or ‘initial’

class autots.GeneralTransformer(outlier_method: str = None, outlier_threshold: float = 3, outlier_position: str = 'first', fillna: str = 'ffill', transformation: str = None, second_transformation: str = None, transformation_param: str = None, detrend: str = None, third_transformation: str = None, transformation_param2: str = None, fourth_transformation: str = None, discretization: str = 'center', n_bins: int = None, coerce_integer: bool = False, grouping: str = None, reconciliation: str = None, grouping_ids=None, constraint=None, random_seed: int = 2020)

Bases: object

Remove outliers, fillNA, then mathematical transformations.

Expects a chronologically sorted pandas.DataFrame with a DatetimeIndex, only numeric data, and a ‘wide’ (one column per series) shape.

Warning

  • inverse_transform will not fully return the original data under some conditions
    • outliers removed or clipped will be returned in the clipped or filled na form

    • NAs filled will be returned with the filled value

    • Discretization cannot be inversed

    • RollingMean, PctChange, CumSum, and DifferencedTransformer will only return original or an immediately following forecast
      • by default ‘forecast’ is expected, ‘original’ can be set in trans_method

Parameters
  • outlier_method (str) –

    • level of outlier removal, if any, per series

    ’None’ ‘clip’ - replace outliers with the highest value allowed by threshold ‘remove’ - remove outliers and replace with np.nan

  • outlier_threshold (float) – number of std deviations from mean to consider an outlier. Default 3.

  • outlier_position (str) – when to remove outliers ‘first’ - remove outliers before other transformations ‘middle’ - remove outliers after first_transformation ‘last’ - remove outliers after fourth_transformation

  • fillNA (str) –

    • method to fill NA, passed through to FillNA()

    ’ffill’ - fill most recent non-na value forward until another non-na value is reached ‘zero’ - fill with zero. Useful for sales and other data where NA does usually mean $0. ‘mean’ - fill all missing values with the series’ overall average value ‘median’ - fill all missing values with the series’ overall median value ‘rolling mean’ - fill with last n (window = 10) values ‘ffill mean biased’ - simple avg of ffill and mean ‘fake date’ - shifts forward data over nan, thus values will have incorrect timestamps

  • transformation (str) –

    • transformation to apply

    ’None’ ‘MinMaxScaler’ - Sklearn MinMaxScaler ‘PowerTransformer’ - Sklearn PowerTransformer ‘QuantileTransformer’ - Sklearn ‘MaxAbsScaler’ - Sklearn ‘StandardScaler’ - Sklearn ‘RobustScaler’ - Sklearn ‘PCA, ‘FastICA’ - performs sklearn decomposition and returns n-cols worth of n_components ‘Detrend’ - fit then remove a linear regression from the data ‘RollingMean’ - 10 period rolling average, can receive a custom window by transformation_param if used as second_transformation ‘FixedRollingMean’ - same as RollingMean, but with inverse_transform disabled, so smoothed forecasts are maintained. ‘RollingMean10’ - 10 period rolling average (smoothing) ‘RollingMean100thN’ - Rolling mean of periods of len(train)/100 (minimum 2) ‘DifferencedTransformer’ - makes each value the difference of that value and the previous value ‘PctChangeTransformer’ - converts to pct_change, not recommended if lots of zeroes in data ‘SinTrend’ - removes a sin trend (fitted to each column) from the data ‘CumSumTransformer’ - makes value sum of all previous ‘PositiveShift’ - makes all values >= 1 ‘Log’ - log transform (uses PositiveShift first as necessary) ‘IntermittentOccurrence’ - -1, 1 for non median values ‘SeasonalDifference’ - remove the last lag values from all values ‘SeasonalDifferenceMean’ - remove the average lag values from all ‘SeasonalDifference7’,’12’,’28’ - non-parameterized version of Seasonal

  • second_transformation (str) – second transformation to apply. Same options as transformation, but with transformation_param passed in if used

  • detrend (str) – Model and remove a linear component from the data. None, ‘Linear’, ‘Poisson’, ‘Tweedie’, ‘Gamma’, ‘RANSAC’, ‘ARD’

  • second_transformation – second transformation to apply. Same options as transformation, but with transformation_param passed in if used

  • transformation_param (str) – passed to second_transformation, not used by most transformers.

  • fourth_transformation (str) – third transformation to apply. Sames options as transformation.

  • discretization (str) – method of binning to apply None - no discretization ‘center’ - values are rounded to center value of each bin ‘lower’ - values are rounded to lower range of closest bin ‘upper’ - values are rounded up to upper edge of closest bin ‘sklearn-quantile’, ‘sklearn-uniform’, ‘sklearn-kmeans’ - sklearn kbins discretizer

  • n_bins (int) – number of quantile bins to split data into

  • coerce_integer (bool) – whether to force inverse_transform into integers

  • random_seed (int) – random state passed through where applicable

fill_na(df, window: int = 10)
Parameters
  • df (pandas.DataFrame) – Datetime Indexed

  • window (int) – passed through to rolling mean fill technique

Returns

pandas.DataFrame

fit(df)

Apply transformations and return transformer object.

Parameters

df (pandas.DataFrame) – Datetime Indexed

fit_transform(df)

Directly fit and apply transformations to convert df.

inverse_transform(df, trans_method: str = 'forecast')

Undo the madness.

Parameters
  • df (pandas.DataFrame) – Datetime Indexed

  • trans_method (str) – ‘forecast’ or ‘original’ passed through

outlier_treatment(df)
Parameters

df (pandas.DataFrame) – Datetime Indexed

Returns

pandas.DataFrame

transform(df)

Apply transformations to convert df.

autots.long_to_wide(df, date_col: str = 'datetime', value_col: str = 'value', id_col: str = 'series_id', frequency: str = 'infer', na_tolerance: float = 0.999, drop_data_older_than_periods: int = 100000, drop_most_recent: int = 0, aggfunc: str = 'first', verbose: int = 1)

Take long data and convert into wide, cleaner data.

Parameters
  • df (pd.DataFrame) –

  • date_col (str) –

  • value_col (str) –

    • the name of the column with the values of the time series (ie sales $)

  • id_col (str) –

    • name of the id column, unique for each time series

  • frequency (str) –

    • frequency in string of alias for DateOffset object, normally “1D” -daily, “MS” -month start etc.

    currently, aliases are listed somewhere in here: https://pandas.pydata.org/pandas-docs/stable/user_guide/timeseries.html

  • na_tolerance (float) –

    • allow up to this percent of values to be NaN, else drop the entire series

    the default of 0.95 means a series can be 95% NaN values and still be included.

  • drop_data_older_than_periods (int) –

    • cut off older data because eventually you just get too much

    10,000 is meant to be rather high, normally for daily data I’d use only the last couple of years, say 1500 samples

  • drop_most_recent (int) –

    • if to drop the most recent data point

    useful if you pull monthly data before month end, and you don’t want an incomplete month appearing complete

  • aggfunc (str) –

    • passed to pd.pivot_table, determines how to aggregate duplicates for series_id and datetime

    other options include “mean” and other numpy functions, beware data must already be input as numeric type for these to work. if categorical data is provided, aggfunc=’first’ is recommended