CODEWITHSUNDEEP
pythontime-seriesanalyticsdata-sciencearima
Pranav Shankar
Pranav Shankar

Reputation: 51

ARIMA modelling for more than 1 time series

So my data looks like this

           Date1 Date2 Date 3.....Date N
Instance1
Instance2
.
.
.

I don't want to build an ARIMA model for only Instance1. I want a universal model that takes all the instances into consideration. I found a lot of examples that show me how to fit it for

         Date1 Date2 Date 3.....Date N
Instance1

but none for all instances

Upvotes: 0

Views: 2085

Answers (1)

Data Junki
Data Junki

Reputation: 164

If you believe your time-series are correlated and you want to take those correlations into account in your forecast/simulation, you should look at vector auto-regression models (VAR). Here are a couple options in python:

StatsModels

PyFlux

If you don't believe they are correlated, then there's no reason you can't just loop through each time-series and apply an ARIMA model one at a time.

Upvotes: 4

Related Questions