Skip to content

Tuning forecaster

Skforecast library allows to combine grid search strategy with backtesting in order to identify the combination of lags and hyperparameters that achieve the best prediction performance.

Libraries

1
2
3
4
import numpy as np
import pandas as pd
from skforecast.ForecasterAutoreg import ForecasterAutoreg
from sklearn.ensembre import Ridge

Data

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
# Download data
# ==============================================================================
url = ('https://raw.githubusercontent.com/JoaquinAmatRodrigo/skforecast/master/data/h2o.csv')
data = pd.read_csv(url, sep=',', header=0, names=['y', 'datetime'])

# Data preprocessing
# ==============================================================================
data['datetime'] = pd.to_datetime(data['datetime'], format='%Y/%m/%d')
data = data.set_index('datetime')
data = data.asfreq('MS')
data = data['y']
data = data.sort_index()

# Split train-test
# ==============================================================================
steps = 24
data_train = data.loc[: '2001-01-01']
data_val = data.loc['2001-01-01' : '2006-01-01']
data_test  = data.loc['2006-01-01':]

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
# Grid search hyperparameter and lags
# ==============================================================================
forecaster = ForecasterAutoreg(
                regressor = RandomForestRegressor(random_state=123),
                lags      = 12 # Placeholder, the value will be overwritten
             )

# Regressor hyperparameters
param_grid = {'n_estimators': [50, 100],
              'max_depth': [5, 10, 15]}

# Lags used as predictors
lags_grid = [3, 10, [1, 2, 3, 20]]

results_grid = grid_search_forecaster(
                        forecaster  = forecaster,
                        y           = data.loc[:'2006-01-01'],
                        param_grid  = param_grid,
                        lags_grid   = lags_grid,
                        steps       = 12,
                        refit       = True,
                        metric      = 'mean_squared_error',
                        initial_train_size = len(data_train),
                        return_best = True,
                        verbose     = False
                    )

1
results_grid
lags params metric max_depth n_estimators
[ 1 2 3 4 5 6 7 8 9 10] {'max_depth': 5, 'n_estimators': 50} 0.0334486 5 50
[ 1 2 3 4 5 6 7 8 9 10] {'max_depth': 10, 'n_estimators': 50} 0.0392212 10 50
[ 1 2 3 4 5 6 7 8 9 10] {'max_depth': 15, 'n_estimators': 100} 0.0392658 15 100
[ 1 2 3 4 5 6 7 8 9 10] {'max_depth': 5, 'n_estimators': 100} 0.0395258 5 100
[ 1 2 3 4 5 6 7 8 9 10] {'max_depth': 10, 'n_estimators': 100} 0.0402408 10 100
[ 1 2 3 4 5 6 7 8 9 10] {'max_depth': 15, 'n_estimators': 50} 0.0407645 15 50
[ 1 2 3 20] {'max_depth': 15, 'n_estimators': 100} 0.0439092 15 100
[ 1 2 3 20] {'max_depth': 5, 'n_estimators': 100} 0.0449923 5 100
[ 1 2 3 20] {'max_depth': 5, 'n_estimators': 50} 0.0462237 5 50
[1 2 3] {'max_depth': 5, 'n_estimators': 50} 0.0486662 5 50
[ 1 2 3 20] {'max_depth': 10, 'n_estimators': 100} 0.0489914 10 100
[ 1 2 3 20] {'max_depth': 10, 'n_estimators': 50} 0.0501932 10 50
[1 2 3] {'max_depth': 15, 'n_estimators': 100} 0.0505563 15 100
[ 1 2 3 20] {'max_depth': 15, 'n_estimators': 50} 0.0512172 15 50
[1 2 3] {'max_depth': 5, 'n_estimators': 100} 0.0531229 5 100
[1 2 3] {'max_depth': 15, 'n_estimators': 50} 0.0602604 15 50
[1 2 3] {'max_depth': 10, 'n_estimators': 50} 0.0609513 10 50
[1 2 3] {'max_depth': 10, 'n_estimators': 100} 0.0673343 10 100

If argument return_best = True, the forecaster is retrained using all data available and the best combination of lags and hyperparameters.

1
skforecast

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
========================================================================== 
<class 'skforecast.ForecasterAutoreg.ForecasterAutoreg.ForecasterAutoreg'> 
========================================================================== 
Regressor: RandomForestRegressor(max_depth=5, n_estimators=50, random_state=123) 
Lags: [ 1  2  3  4  5  6  7  8  9 10] 
Window size: 10 
Included exogenous: False 
Type of exogenous variable: None 
Exogenous variables names: None 
Training range: [Timestamp('1991-07-01 00:00:00'), Timestamp('2006-01-01 00:00:00')] 
Training index type: <class 'pandas.core.indexes.datetimes.DatetimeIndex'> 
Training index frequancy: MS 
Regressor parameters: {'bootstrap': True, 'ccp_alpha': 0.0, 'criterion': 'squared_error', 'max_depth': 5, 'max_features': 'auto', 'max_leaf_nodes': None, 'max_samples': None, 'min_impurity_decrease': 0.0, 'min_samples_leaf': 1, 'min_samples_split': 2, 'min_weight_fraction_leaf': 0.0, 'n_estimators': 50, 'n_jobs': None, 'oob_score': False, 'random_state': 123, 'verbose': 0, 'warm_start': False}