Time Series Forecasting — Parallel Processing using Pandas Function APIs
In our previous blog on Time Series forecasting, we’ve seen how to use various Univariate forecasting methods to predict a time series dataset. In real life, we may need to run prediction algorithms on multiple similar datasets at the same time i.e. in batch. We can run the forecasting algorithm sequentially but, that will take long time based each dataset volume and number of datasets in a batch.
In this blog, we’ll see how we can use Pandas Function API to run forecasting algorithm(s) on multiple datasets parallelly (in a batch).
Dataset
We have taken a weather dataset — average temperatures across all of the climate stations in a country; we have selected three countries with few years of data (no. of data points: UK = 118, Norway = 106 and Australia = 118).
We want to predict average temperatures for further 18 months.
