Advanced Prophet Usage

You can scan through prophet 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

[1]:

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

[2]:

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

[3]:

from hcrystalball.wrappers import ProphetWrapper

[4]:

ProphetWrapper?

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

[5]:

extra_holidays = {
    'Christmas Day':{'lower_window': -2, 'upper_window':2, 'prior_scale': 20},
#     'Good Friday':{'lower_window': -1, 'upper_window':1, 'prior_scale': 30}
}

Unusual Seasonalities

[6]:

extra_seasonalities = [
    {
        'name':'bi-weekly',
        'period': 14.,
        'fourier_order': 5,
        'prior_scale': 10.0,
        'mode': None
    },
    {
        'name':'bi-yearly',
        'period': 365*2.,
        'fourier_order': 5,
        'prior_scale': 5.0,
        'mode': None
    },
]

Exogenous Variables

[7]:

from sklearn.pipeline import Pipeline
from hcrystalball.feature_extraction import HolidayTransformer

[8]:

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

[9]:

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

[10]:

pipeline = Pipeline([
    ('holidays_de', HolidayTransformer(country_code = 'DE')),
    ('model', prophet)
])
pipeline_extra = Pipeline([
    ('holidays_de', HolidayTransformer(country_code = 'DE')),
    ('model', prophet_extra)
])

[11]:

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

prds.plot(title=f"MAE:{(prds['target']-prds['prophet']).abs().mean().round(3)}");

[11]:

<AxesSubplot:title={'center':'MAE:1.988'}>
../../../_images/examples_tutorial_wrappers_06_advanced_prophet_14_1.png 
[12]:

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

prds_extra.plot(title=f"MAE:{(prds_extra['target']-prds_extra['prophet_extra']).abs().mean().round(3)}");

[12]:

<AxesSubplot:title={'center':'MAE:3.716'}>
../../../_images/examples_tutorial_wrappers_06_advanced_prophet_15_1.png

Compared to non-tweaked model, we are able to better catch the series dynamics, but don’t win in MAE against roughly average predictions

[13]:

prds = (ProphetWrapper().fit(X[:-10], y[:-10])
                 .predict(X[-10:])
                 .merge(y, left_index=True, right_index=True, how='outer')
                 .tail(50)
)
prds.plot(title=f"MAE:{(prds['target']-prds['prophet']).abs().mean().round(3)}");

[13]:

<AxesSubplot:title={'center':'MAE:1.886'}>
../../../_images/examples_tutorial_wrappers_06_advanced_prophet_17_1.png

Full Prophet Output

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

[14]:

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

[14]:
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 6.151973 1.038690 11.545119 6.151973 6.151973 -0.031628 -0.031628 -0.031628 -0.031628 -0.031628 -0.031628 0.0 0.0 0.0 6.120344
2018-12-23 6.149407 1.087590 11.920487 6.149407 6.149407 -0.084270 -0.084270 -0.084270 -0.084270 -0.084270 -0.084270 0.0 0.0 0.0 6.065137
2018-12-24 6.146841 0.857559 11.610195 6.146841 6.146841 -0.010959 -0.010959 -0.010959 -0.010959 -0.010959 -0.010959 0.0 0.0 0.0 6.135883
2018-12-25 6.144276 0.756582 11.534445 6.144276 6.144276 0.029113 0.029113 0.029113 0.029113 0.029113 0.029113 0.0 0.0 0.0 6.173389
2018-12-26 6.141710 0.779932 11.676427 6.141710 6.141710 0.045507 0.045507 0.045507 0.045507 0.045507 0.045507 0.0 0.0 0.0 6.187217
2018-12-27 6.139144 0.594331 11.537535 6.139144 6.139144 0.009433 0.009433 0.009433 0.009433 0.009433 0.009433 0.0 0.0 0.0 6.148578
2018-12-28 6.136579 0.937165 11.405285 6.136578 6.136590 0.042803 0.042803 0.042803 0.042803 0.042803 0.042803 0.0 0.0 0.0 6.179382
2018-12-29 6.134013 0.000953 11.264760 6.133966 6.134073 -0.031628 -0.031628 -0.031628 -0.031628 -0.031628 -0.031628 0.0 0.0 0.0 6.102385
2018-12-30 6.131447 0.426198 11.261615 6.131264 6.131579 -0.084270 -0.084270 -0.084270 -0.084270 -0.084270 -0.084270 0.0 0.0 0.0 6.047178
2018-12-31 6.128882 0.899522 11.434409 6.128583 6.129082 -0.010959 -0.010959 -0.010959 -0.010959 -0.010959 -0.010959 0.0 0.0 0.0 6.117923