3
$\begingroup$

I have a table with the following columns:

Date(Month,Year), Sold_Past_Month, Quantity_Available, Quantity_Shipping_In, Missed_Sales, Quantity_Needed

Quantity_Needed is the dependent variable that is numerical.

I want to predict Quantity_Needed and train a model using the columns mentioned above; however, I also want Quantity_Needed to train the model.

I have data from the past 5 years -> so I would have roughly 60 rows of data for one item.

Is it possible to use ARIMA for this?

If it is, what should I do next to build my model?

|improve this question|||||
$\endgroup$
  • 1
    $\begingroup$ Hey! Welcome! Yes, ARIMA is actually constructed with time series in mind. If you are going to be successful is something you will only know if you try. You can go trough this tutorial for creating an ARIMA model with sklearn in Python. $\endgroup$ – Pedro Henrique Monforte Apr 9 '19 at 23:39
0
$\begingroup$

The correct name of what you are looking for is: ARIMAX, the ARIMA model works without covariables, the X (exogenous variables) include the possibility of having other explanatory variables.

https://robjhyndman.com/hyndsight/arimax/

|improve this answer|||||
$\endgroup$
0
$\begingroup$

I am not sure about the nature of this problem because I see that even the independent variables are linked to their past values, if so and if you don’t have a test data set you can try VAR model.

In a VAR model, each variable is a linear function of the past values of itself and the past values of all the other variables.

Otherwise you can also use LSTMs.

|improve this answer|||||
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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