I have a dataset consisting of purchasing history from an e-commerce website. The columns consist of customer id, product id, postal code, quantity of products purchased, date of the purchase. There are thousands of different customer ids, hundreds of product ids, and some million rows.

I ran ARIMA modelling for forecasting the purchased quantities of a given product. Now I want to try other methods for analysing the dataset, but cannot find models which could fit the dataset well. What other models could I run in order to gain more insight on this data?

  • $\begingroup$ Are you familiar with neural networks? $\endgroup$ – Media Sep 7 '18 at 15:52
  • $\begingroup$ @Media I am a bit $\endgroup$ – Snow Sep 7 '18 at 17:14
  • $\begingroup$ Please specify the input and output spaces. $\endgroup$ – Media Sep 8 '18 at 9:48
  • $\begingroup$ @Media for example predicting customer churn based on the previous purchases of the customer. I don't really know which of the dataset columns I could use for input. I would guess first I'd need to map the categorical values to integers. $\endgroup$ – Snow Sep 10 '18 at 10:23

Based on your dataset you can also try to create a product recommendation model so that given a new customer that just purchased a particular item you can predict/suggest or recommend what other items they could purchase next during their shopping trip or after: - Collaborative filtering can be used to achieve this - Frequent pattern mining can also be used.


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.