I am trying to do a side project to get a better understanding of the whole data science after completing my online course. Am now in an early stage of just laying out the project in general and was wondering if you can put time series data and just variable (normalized if necessary) in a multivariable polynomial regression model?

  • $\begingroup$ @Spectre answered it perfectly for you. But I would like to add, for your future understanding, that "regular" vs. time series data is inherently different. This distinction is known as "Cross-sectional" vs. "Longitudinal". Cross sectional data is observations of events for multiple subjects at "snapshots" in time. Longitudinal data is observations of events over some period of time for one subject. This distinction is why time series models are inherently different from "typical" models. $\endgroup$ May 12 at 0:03
  • $\begingroup$ @AndrewJaeyoung Thank you for the additional information! My data is a traffic data taken for a certain period of time (15min) and has the information such as the date, time, type of day, average speed of vehicles, overall traffic flow. So, I guess this qualifies as a cross-sectional data? I've been looking through online and can't seem to find a entry level model that takes all of that data into account. Does it have to be NN model? $\endgroup$
    – ENGSSG
    May 12 at 9:57

1 Answer 1


Can you practically use Polynomial Regression on time series data? YES YOU CAN!

But that does not mean you should!

Time series data and non time series data are 2 very different kind. The models which can work on time and non time series data are also different. Even the preprocessing techniques used in non time series data cannot be simple used in time series data.

The reason for this is that there is an inherent order/sequence present in time series datasets whereas non time series datasets do not have that sequence.

Polynomial Regression is usually used in non time series data. You can try training it on your time series dataset to see what kind of results you get. These kinds of models do not take into account the sequence/order into account so I would not hope for much.

Instead you could use ARIMA, SARIMA, SARIMAX, VAR, VARMA, VARMAX, HWES or more complex Deep Learning models like RNN's, LSTM's and more. They would be much more appropriate results.

  • 1
    $\begingroup$ Thank you so much! I had an off feeling why it would not work, but your explanation clearly explains why I should not use it. I'll look into the models that you have mentioned. :) Now time to study :D $\endgroup$
    – ENGSSG
    May 11 at 19:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.