0
$\begingroup$

I have a panel data which consists of the sales of different companies across different years. Can someone tell me what kind of regression and ML methods we can use for forecasting the sales?

$\endgroup$

2 Answers 2

1
$\begingroup$

Sales data is Time-series data. You should use the method of Time-series forecasting.

Below is a list of useful links to start with -

Analytics Vidhya

Yashuseth blog

Kaggle blog

This is a good book to grasp the concepts -
Forecasting: Principles and Practice

Also, do check this similar question.
Stack Exchange

$\endgroup$
1
$\begingroup$

So I have interpreted the question as asking the following (let me know if this is what you have in mind): what prediction methods/models we can utilise for predict future sales from historical sales data for different companies?

So, in this exercise, it is always essential to define a baseline. With a baseline, we can then experiment with other models to understand if they perform better or worse. Therefore, as a baseline (depending on how the data is structured), you can format this into a regression model, where the input is a matrix with $N$ rows and $T$ columns, where $t \in T$ represents a particular time point (if the data is structured by dates and the sales data across all companies is taken at the same time points).

If that is not the case, then we need to look towards a more complex input data structure, where we need to include the time point $t$, as well as the value at that time point $t$. Therefore, we could use for example a Recurrent Neural Network or a Long-Short Term Memory neural network to predict the future sale values given the historical sales data (as baseline models). Hope this helps and good luck!

$\endgroup$

Your Answer

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

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