Hands-On Machine Learning on Google Cloud Platform

Multiplicative method

As we said, the seasonal_decompose() performs both additive and multiplicative decomposition. To run multiplicative method, just type the following command:

DecompDataMult = seasonal_decompose(data, model='multiplicative')

At this point, we just have to visualize the result:

DecompDataMult.plot()plt.show()

The following graph shows the decomposition results by multiplicative method:

In the previous figure, we can note that the trend and seasonality information extracted from the time series do seem reasonable. The residuals show an interesting variation; periods of high variability are clearly identified in the early ...