Feature importance¶

The importance of the predictors included in a forecaster can be obtained using the method get_feature_importance. This method accesses the coef_ and feature_importances_ attributes of the internal regressor.

  Warning

This method only returns values if the regressor used inside the forecaster has the attribute `coef_` or `feature_importances_`.

Libraries¶

In [1]:

# Libraries
# ==============================================================================
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.ensemble import RandomForestRegressor
from sklearn.linear_model import Ridge
from skforecast.ForecasterAutoreg import ForecasterAutoreg
from skforecast.ForecasterAutoregDirect import ForecasterAutoregDirect

# Libraries # ============================================================================== import pandas as pd import matplotlib.pyplot as plt from sklearn.ensemble import RandomForestRegressor from sklearn.linear_model import Ridge from skforecast.ForecasterAutoreg import ForecasterAutoreg from skforecast.ForecasterAutoregDirect import ForecasterAutoregDirect

Data¶

In [2]:

# Download data
# ==============================================================================
url = ('https://raw.githubusercontent.com/JoaquinAmatRodrigo/skforecast/master/data/h2o_exog.csv')
data = pd.read_csv(url, sep=',', header=0, names=['date', 'y', 'exog_1', 'exog_2'])

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

# Download data # ============================================================================== url = ('https://raw.githubusercontent.com/JoaquinAmatRodrigo/skforecast/master/data/h2o_exog.csv') data = pd.read_csv(url, sep=',', header=0, names=['date', 'y', 'exog_1', 'exog_2']) # Data preprocessing # ============================================================================== data['date'] = pd.to_datetime(data['date'], format='%Y/%m/%d') data = data.set_index('date') data = data.asfreq('MS')

Extract feature importance from trained forecaster¶

In [3]:

# Create and fit forecaster
# ==============================================================================
forecaster = ForecasterAutoreg(
                 regressor = RandomForestRegressor(),
                 lags = 5
             )

forecaster.fit(y=data['y'], exog=data[['exog_1', 'exog_2']])

# Predictors importance
# ==============================================================================
forecaster.get_feature_importance()

# Create and fit forecaster # ============================================================================== forecaster = ForecasterAutoreg( regressor = RandomForestRegressor(), lags = 5 ) forecaster.fit(y=data['y'], exog=data[['exog_1', 'exog_2']]) # Predictors importance # ============================================================================== forecaster.get_feature_importance()

Out[3]:

	feature	importance
0	lag_1	0.541886
1	lag_2	0.090949
2	lag_3	0.023549
3	lag_4	0.073510
4	lag_5	0.063987
5	exog_1	0.045629
6	exog_2	0.160490

In [4]:

# Create and fit forecaster
# ==============================================================================
forecaster = ForecasterAutoreg(
                 regressor = Ridge(),
                 lags = 5
             )

forecaster.fit(y=data['y'], exog=data[['exog_1', 'exog_2']])

forecaster.get_feature_importance()

# Create and fit forecaster # ============================================================================== forecaster = ForecasterAutoreg( regressor = Ridge(), lags = 5 ) forecaster.fit(y=data['y'], exog=data[['exog_1', 'exog_2']]) forecaster.get_feature_importance()

Out[4]:

	feature	importance
0	lag_1	0.327688
1	lag_2	-0.073593
2	lag_3	-0.152202
3	lag_4	-0.217106
4	lag_5	-0.145800
5	exog_1	0.379798
6	exog_2	0.668162

Since ForecasterAutoregDirect fits one model per step, it is necessary to specify from which model retrieves its feature importance.

In [5]:

# Create and fit forecaster
# ==============================================================================
forecaster = ForecasterAutoregDirect(
                 regressor = RandomForestRegressor(),
                 steps = 10,
                 lags = 5
             )

forecaster.fit(y=data['y'], exog=data[['exog_1', 'exog_2']])

# Predictors importance
# ==============================================================================
forecaster.get_feature_importance(step=1)

# Create and fit forecaster # ============================================================================== forecaster = ForecasterAutoregDirect( regressor = RandomForestRegressor(), steps = 10, lags = 5 ) forecaster.fit(y=data['y'], exog=data[['exog_1', 'exog_2']]) # Predictors importance # ============================================================================== forecaster.get_feature_importance(step=1)

Out[5]:

	feature	importance
0	lag_1	0.521185
1	lag_2	0.098278
2	lag_3	0.026642
3	lag_4	0.081040
4	lag_5	0.046787
5	exog_1	0.046945
6	exog_2	0.179123

In [6]:

%%html
<style>
.jupyter-wrapper .jp-CodeCell .jp-Cell-inputWrapper .jp-InputPrompt {display: none;}
</style>

%%html