mirror of
https://github.com/saymrwulf/prophet.git
synced 2026-05-15 20:50:48 +00:00
79d0793ce4
* Resolve conflict * Change comments and add error column to output DataFrame * Change file structure * Update * Modified diagnostics * Update diagnostics.py following the advice on Github * Add tests and documentation * Change copy method into Prophet class and reflect comments
88 lines
4 KiB
Python
88 lines
4 KiB
Python
# Copyright (c) 2017-present, Facebook, Inc.
|
|
# All rights reserved.
|
|
#
|
|
# This source code is licensed under the BSD-style license found in the
|
|
# LICENSE file in the root directory of this source tree. An additional grant
|
|
# of patent rights can be found in the PATENTS file in the same directory.
|
|
|
|
from __future__ import absolute_import
|
|
from __future__ import division
|
|
from __future__ import print_function
|
|
from __future__ import unicode_literals
|
|
|
|
import os
|
|
import numpy as np
|
|
import pandas as pd
|
|
|
|
from unittest import TestCase
|
|
from fbprophet import Prophet
|
|
from fbprophet import diagnostics
|
|
|
|
|
|
class TestDiagnostics(TestCase):
|
|
|
|
def __init__(self, *args, **kwargs):
|
|
super(TestDiagnostics, self).__init__(*args, **kwargs)
|
|
# Use first 100 record in data.csv
|
|
self.__df = pd.read_csv(os.path.join(os.path.dirname(__file__), 'data.csv'), parse_dates=['ds']).head(100)
|
|
|
|
def test_simulated_historical_forecasts(self):
|
|
m = Prophet()
|
|
m.fit(self.__df)
|
|
k = 3
|
|
for p in [1, 10]:
|
|
for h in [1, 3]:
|
|
period = '{} days'.format(p)
|
|
horizon = '{} days'.format(h)
|
|
df_shf = diagnostics.simulated_historical_forecasts(m, horizon=horizon, k=k, period=period)
|
|
# All cutoff dates should be less than ds dates
|
|
self.assertTrue((df_shf['cutoff'] < df_shf['ds']).all())
|
|
# The unique size of output cutoff should be equal to 'k'
|
|
self.assertEqual(len(np.unique(df_shf['cutoff'])), k)
|
|
# Each y in df_shf and self.__df with same ds should be equal
|
|
df_merged = pd.merge(df_shf, self.__df, 'left', on='ds')
|
|
self.assertAlmostEqual(np.sum((df_merged['y_x'] - df_merged['y_y']) ** 2), 0.0)
|
|
|
|
def test_simulated_historical_forecasts_logistic(self):
|
|
m = Prophet(growth='logistic')
|
|
df = self.__df.copy()
|
|
df['cap'] = 40
|
|
m.fit(df)
|
|
df_shf = diagnostics.simulated_historical_forecasts(m, horizon='3 days', k=2, period='3 days')
|
|
# All cutoff dates should be less than ds dates
|
|
self.assertTrue((df_shf['cutoff'] < df_shf['ds']).all())
|
|
# The unique size of output cutoff should be equal to 'k'
|
|
self.assertEqual(len(np.unique(df_shf['cutoff'])), 2)
|
|
# Each y in df_shf and self.__df with same ds should be equal
|
|
df_merged = pd.merge(df_shf, df, 'left', on='ds')
|
|
self.assertAlmostEqual(np.sum((df_merged['y_x'] - df_merged['y_y']) ** 2), 0.0)
|
|
|
|
def test_simulated_historical_forecasts_default_value_check(self):
|
|
m = Prophet()
|
|
m.fit(self.__df)
|
|
# Default value of period should be equal to 0.5 * horizon
|
|
df_shf1 = diagnostics.simulated_historical_forecasts(m, horizon='10 days', k=1)
|
|
df_shf2 = diagnostics.simulated_historical_forecasts(m, horizon='10 days', k=1, period='5 days')
|
|
self.assertAlmostEqual(((df_shf1 - df_shf2)**2)[['y', 'yhat']].sum().sum(), 0.0)
|
|
|
|
def test_cross_validation(self):
|
|
m = Prophet()
|
|
m.fit(self.__df)
|
|
# Calculate the number of cutoff points(k)
|
|
te = self.__df['ds'].max()
|
|
ts = self.__df['ds'].min()
|
|
horizon = pd.Timedelta('4 days')
|
|
period = pd.Timedelta('1 days')
|
|
initial = pd.Timedelta('90 days')
|
|
k = int(np.floor(((te - horizon) - (ts + initial)) / period))
|
|
df_cv = diagnostics.cross_validation(m, horizon=horizon, period=period, initial=initial)
|
|
# The unique size of output cutoff should be equal to 'k'
|
|
self.assertEqual(len(np.unique(df_cv['cutoff'])), k)
|
|
|
|
def test_cross_validation_default_value_check(self):
|
|
m = Prophet()
|
|
m.fit(self.__df)
|
|
# Default value of initial should be equal to 3 * horizon
|
|
df_cv1 = diagnostics.cross_validation(m, horizon='32 days', period='1 days')
|
|
df_cv2 = diagnostics.cross_validation(m, horizon='32 days', period='1 days', initial='96 days')
|
|
self.assertAlmostEqual(((df_cv1 - df_cv2)**2)[['y', 'yhat']].sum().sum(), 0.0)
|