Forecasting Economic Time Series in Python using SARIMAX

preview_player
Показать описание
In this tutorial we show how to forecast vehicle sales data into the future using the SARIMAX model from Python's Statsmodels package.

We cover the following topics in this tutorial:
1) Imputing missing values in time series
2) Forecasting with multiple series i.e. exogenous regressors
3) Time series cross validation
4) Hyperparameter selection using randomized stochastic optimization algorithms

Code and data can be found at my GitHub below:
Рекомендации по теме
Комментарии
Автор

This is genius. Thank you so much. Most helpful thing I've found on the internet, as it pertains to Sarimax.

GraceHencey
Автор

Great stuff. Very clean teaching style.

eervin
Автор

Very cool explanation, dude! Thank you

mateusacorrea
Автор

Great tutorial. Any books you recommend?

jnicune
Автор

Wonderful tutorial, thank you so much! Just one question: To make the code more succinct, at 31:45 where you link the model to the score function and define the order, trend, and seasonal variable and index the hyperparameters, couldn't you index the hyperparameters once in a list instead, i.e., params[, , ] instead of params [] 3x?

adrianfischbach
Автор

Very well done. Thank you very much @datamodelingxp, Michael. Have compared your method to auto_arima model?

AbdallahIDRISSE
welcome to shbcf.ru