Simple auto_arima model

This is a simple example of how we can fit an ARIMA model in several lines without knowing anything about our data or optimal hyper parameters.

Out:

Performing stepwise search to minimize aic
 ARIMA(2,1,2)(1,0,1)[12] intercept   : AIC=2915.641, Time=8.22 sec
 ARIMA(0,1,0)(0,0,0)[12] intercept   : AIC=3049.597, Time=0.09 sec
 ARIMA(1,1,0)(1,0,0)[12] intercept   : AIC=2954.973, Time=3.42 sec
 ARIMA(0,1,1)(0,0,1)[12] intercept   : AIC=2947.014, Time=4.20 sec
 ARIMA(0,1,0)(0,0,0)[12]             : AIC=3047.612, Time=0.07 sec
 ARIMA(2,1,2)(0,0,1)[12] intercept   : AIC=2938.457, Time=7.53 sec
 ARIMA(2,1,2)(1,0,0)[12] intercept   : AIC=2915.636, Time=6.76 sec
 ARIMA(2,1,2)(0,0,0)[12] intercept   : AIC=2984.606, Time=0.15 sec
 ARIMA(2,1,2)(2,0,0)[12] intercept   : AIC=2915.168, Time=21.95 sec
 ARIMA(2,1,2)(2,0,1)[12] intercept   : AIC=2917.256, Time=28.60 sec
 ARIMA(1,1,2)(2,0,0)[12] intercept   : AIC=2924.379, Time=25.33 sec
 ARIMA(2,1,1)(2,0,0)[12] intercept   : AIC=2911.007, Time=24.57 sec
 ARIMA(2,1,1)(1,0,0)[12] intercept   : AIC=2911.505, Time=7.82 sec
 ARIMA(2,1,1)(2,0,1)[12] intercept   : AIC=2912.933, Time=23.59 sec
 ARIMA(2,1,1)(1,0,1)[12] intercept   : AIC=2911.904, Time=7.20 sec
 ARIMA(1,1,1)(2,0,0)[12] intercept   : AIC=2919.168, Time=18.49 sec
 ARIMA(2,1,0)(2,0,0)[12] intercept   : AIC=2927.636, Time=23.31 sec
 ARIMA(3,1,1)(2,0,0)[12] intercept   : AIC=2912.748, Time=25.20 sec
 ARIMA(1,1,0)(2,0,0)[12] intercept   : AIC=2954.725, Time=20.29 sec
 ARIMA(3,1,0)(2,0,0)[12] intercept   : AIC=2914.169, Time=26.13 sec
 ARIMA(3,1,2)(2,0,0)[12] intercept   : AIC=2915.130, Time=31.39 sec
 ARIMA(2,1,1)(2,0,0)[12]             : AIC=2908.085, Time=21.58 sec
 ARIMA(2,1,1)(1,0,0)[12]             : AIC=2908.791, Time=7.09 sec
 ARIMA(2,1,1)(2,0,1)[12]             : AIC=2909.997, Time=20.50 sec
 ARIMA(2,1,1)(1,0,1)[12]             : AIC=2907.980, Time=6.94 sec
 ARIMA(2,1,1)(0,0,1)[12]             : AIC=2932.875, Time=6.57 sec
 ARIMA(2,1,1)(1,0,2)[12]             : AIC=2909.946, Time=21.90 sec
 ARIMA(2,1,1)(0,0,0)[12]             : AIC=2979.838, Time=0.05 sec
 ARIMA(2,1,1)(0,0,2)[12]             : AIC=2920.616, Time=18.59 sec
 ARIMA(2,1,1)(2,0,2)[12]             : AIC=2911.929, Time=24.35 sec
 ARIMA(1,1,1)(1,0,1)[12]             : AIC=2915.102, Time=5.15 sec
 ARIMA(2,1,0)(1,0,1)[12]             : AIC=2925.280, Time=5.40 sec
 ARIMA(3,1,1)(1,0,1)[12]             : AIC=2910.007, Time=7.98 sec
 ARIMA(2,1,2)(1,0,1)[12]             : AIC=2913.030, Time=8.69 sec
 ARIMA(1,1,0)(1,0,1)[12]             : AIC=2952.472, Time=4.01 sec
 ARIMA(1,1,2)(1,0,1)[12]             : AIC=2922.341, Time=6.39 sec
 ARIMA(3,1,0)(1,0,1)[12]             : AIC=2912.111, Time=7.22 sec
 ARIMA(3,1,2)(1,0,1)[12]             : AIC=2913.230, Time=8.30 sec

Best model:  ARIMA(2,1,1)(1,0,1)[12]
Total fit time: 495.266 seconds

print(__doc__)

# Author: Taylor Smith <taylor.smith@alkaline-ml.com>

import pmdarima as pm
from pmdarima import model_selection
import numpy as np
from matplotlib import pyplot as plt

# #############################################################################
# Load the data and split it into separate pieces
data = pm.datasets.load_wineind()
train, test = model_selection.train_test_split(data, train_size=150)

# Fit a simple auto_arima model
arima = pm.auto_arima(train, error_action='ignore', trace=True,
                      suppress_warnings=True, maxiter=10,
                      seasonal=True, m=12)

# #############################################################################
# Plot actual test vs. forecasts:
x = np.arange(test.shape[0])
plt.scatter(x, test, marker='x')
plt.plot(x, arima.predict(n_periods=test.shape[0]))
plt.title('Actual test samples vs. forecasts')
plt.show()

Total running time of the script: ( 8 minutes 15.379 seconds)