Advanced Prophet Usage

You can scan through fbprophet docs and find many options how to tweak your model. Some of that functionality is moved to initialization stage to be compatible with Sklearn API. We will showcase the parts that were moved to initialization, but you can also look for other model parameters that could help fine-tuning your model

from hcrystalball.wrappers import ProphetWrapper

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
plt.style.use('seaborn')
plt.rcParams['figure.figsize'] = [12, 6]

from hcrystalball.utils import generate_tsdata
X, y = generate_tsdata(n_dates=365*2)

ProphetWrapper?

Init signature:
ProphetWrapper(
    growth='linear',
    changepoints=None,
    n_changepoints=25,
    changepoint_range=0.8,
    yearly_seasonality='auto',
    weekly_seasonality='auto',
    daily_seasonality='auto',
    holidays=None,
    seasonality_mode='additive',
    seasonality_prior_scale=10.0,
    holidays_prior_scale=10.0,
    changepoint_prior_scale=0.05,
    mcmc_samples=0,
    interval_width=0.8,
    uncertainty_samples=1000,
    stan_backend=None,
    name='prophet',
    conf_int=False,
    full_prophet_output=False,
    extra_seasonalities=None,
    extra_regressors=None,
    extra_holidays=None,
    fit_params=None,
    clip_predictions_lower=None,
    clip_predictions_upper=None,
)
Docstring:
Wrapper for fbprophet.Prophet model

https://facebook.github.io/prophet/docs/quick_start.html#python-api

Bring fbprophet to sklearn time-series compatible interface and puts fit parameters
to initialization stage.

Parameters
----------
name : str
    Name of the model instance, used also as column name for returned prediction.

conf_int : bool
    Whether confidence intervals should be also outputed.

full_prophet_output: bool
    Whether the `predict` method should output the full fbprophet.Prophet dataframe.

extra_seasonalities : list of dicts
    Dictionary will be passed to fbprophet.Prophet add_regressor method.

extra_regressors : list or list of dicts
    Dictionary will be passed to fbprophet.Prophet add_seasonality method.

extra_holidays : dict of dict
    Dict with name of the holiday and values as another dict with required
    'lower_window' key and 'upper_window' key and optional 'prior_scale' key
    i.e.{'holiday_name': {'lower_window':1, 'upper_window:1, 'prior_scale: 10}}.

fit_params : dict
    Parameters passed to `fit` fbprophet.Prophet model.

clip_predictions_lower : float
    Minimal value allowed for predictions - predictions will be clipped to this value.

clip_predictions_upper : float
    Maximum value allowed for predictions - predictions will be clipped to this value.
File:           ~/data_science/time-series-framework/src/hcrystalball/wrappers/_prophet.py
Type:           ABCMeta
Subclasses:

Advanced Holidays

For holidays, we are able to define instead of single boolean attribute distribution around given day. We define lower_window, upper_window and prior_scales

extra_holidays = {
    'Whit Monday':{'lower_window': 2, 'upper_window':2, 'prior_scale': 10},
    'Good Friday':{'lower_window': 1, 'upper_window':1, 'prior_scale': 30}
}

Unusual Seasonalities

extra_seasonalities = [
    {
        'name':'bi-weekly',
        'period': 14.,
        'fourier_order': 5,
        'prior_scale': 15.0,
        'mode': None
    },
    {
        'name':'bi-yearly',
        'period': 365*2.,
        'fourier_order': 5,
        'prior_scale': 5.0,
        'mode': None
    },
]

Exogenous Variables

extra_regressors = ['trend_line']
X['trend_line'] = np.arange(len(X))

prophet = ProphetWrapper(
    extra_holidays=extra_holidays,
    extra_seasonalities=extra_seasonalities,
    extra_regressors=extra_regressors,
    name='prophet_extra',
)

(prophet.fit(X[:-10], y[:-10])
        .predict(X[-10:])
        .merge(y, left_index=True, right_index=True, how='outer')
        .tail(50)
        .plot()
);

Compared to non-tweaked model, this shows our fine-tuning helped to increase performance significantly

(ProphetWrapper().fit(X[:-10], y[:-10])
                 .predict(X[-10:])
                 .merge(y, left_index=True, right_index=True, how='outer')
                 .tail(50)
                 .plot()
);

Full Prophet Output

If you need, you can also pass full_prophet_output and get rich predict output

(ProphetWrapper(full_prophet_output=True, conf_int=True)
     .fit(X[:-10], y[:-10])
     .predict(X[-10:])
)

	trend	prophet_lower	prophet_upper	trend_lower	trend_upper	additive_terms	additive_terms_lower	additive_terms_upper	weekly	weekly_lower	weekly_upper	multiplicative_terms	multiplicative_terms_lower	multiplicative_terms_upper	prophet
2018-12-22	6.184323	0.940298	11.284113	6.184323	6.184323	-0.019427	-0.019427	-0.019427	-0.019427	-0.019427	-0.019427	0.0	0.0	0.0	6.164896
2018-12-23	6.182180	1.150368	11.390180	6.182180	6.182180	-0.091169	-0.091169	-0.091169	-0.091169	-0.091169	-0.091169	0.0	0.0	0.0	6.091011
2018-12-24	6.180037	0.787339	11.328600	6.180037	6.180037	-0.010261	-0.010261	-0.010261	-0.010261	-0.010261	-0.010261	0.0	0.0	0.0	6.169776
2018-12-25	6.177894	0.556975	11.361476	6.177894	6.177894	0.020382	0.020382	0.020382	0.020382	0.020382	0.020382	0.0	0.0	0.0	6.198277
2018-12-26	6.175752	1.031746	11.077620	6.175752	6.175752	0.015292	0.015292	0.015292	0.015292	0.015292	0.015292	0.0	0.0	0.0	6.191044
2018-12-27	6.173609	1.011388	11.724293	6.173609	6.173621	0.054739	0.054739	0.054739	0.054739	0.054739	0.054739	0.0	0.0	0.0	6.228348
2018-12-28	6.171466	1.030493	11.290906	6.171466	6.171515	0.030444	0.030444	0.030444	0.030444	0.030444	0.030444	0.0	0.0	0.0	6.201909
2018-12-29	6.169323	1.187374	11.720556	6.169308	6.169430	-0.019427	-0.019427	-0.019427	-0.019427	-0.019427	-0.019427	0.0	0.0	0.0	6.149896
2018-12-30	6.167180	0.381503	11.488146	6.167136	6.167376	-0.091169	-0.091169	-0.091169	-0.091169	-0.091169	-0.091169	0.0	0.0	0.0	6.076011
2018-12-31	6.165037	0.871889	11.621956	6.164944	6.165293	-0.010261	-0.010261	-0.010261	-0.010261	-0.010261	-0.010261	0.0	0.0	0.0	6.154776