TransWikia.com

Forecasting Consumption for Multiple Products for Multiple Regions

Data Science Asked by AatmaRam on November 15, 2020

Came across a very interesting Real-World Time Series Forecast Problem.

Can you please help me understand the right track to resolve the below Time Series problem:

Input Data Sample:
enter image description here

and we want to predict the consumption of Product at each location by type for next 6 months.

e.g.

I have tried doing it by building a simple ARIMA model to get monthly values and then multiplied by the factors calculated for consumption by Product, Type etc.

But can this problem be solved by any other Time Series Technique?
Can we use SARIMAX? If yes, then how can we deal with exogenous regressors being Categorical Variables?

Few observations in data:

  1. Tea might have been consumed by a country in 2017 have now been replaced by Tea2 (Tea type 2) for next years, so how does time series maintain the continuity in data.

Any leads and solution will be highly appreciated.

One Answer

Models

  • Xgboost can be used for Time Series. Here is a tutorial on Kaggle.

  • If Deep Learning is an option (the data set has at least several thousand samples), you can try a RNN (Recurrent Neural Network), with LSTM or GRU cells.

Features
You can transform categorical variables into indicator variables with the get_dummies function in pandas.

Answered by θ Grunberg on November 15, 2020

Add your own answers!

Ask a Question

Get help from others!

© 2024 TransWikia.com. All rights reserved. Sites we Love: PCI Database, UKBizDB, Menu Kuliner, Sharing RPP