Subscribe to DSC Newsletter


I am looking for robust statistical based model that could predict the overall traffic (visits) of a web-site on any future date.

I have visits data on daily basis. The one method that I thought is to use ARIMA based prediction to estimate visits (Time-Series modelling).


Please suggest other methods and what other data might be looking at if we need to predict the website traffic.

Tags: Model, Statistical

Views: 1055

Replies to This Discussion


Hi Yashu


I'd concur and recommend a Time Series approach to start with. You might want to keep it simpler and look at Exponential Smoothing to start and then "move up" to ARIMA after that. ARIMA will allow you to include predictor variables e.g. Campaign Attributes.


That is not to say that more Cross-Sectional modelling techiques e.g. Decision Trees, Neural Nets, etc. would not help. But it sounds like your problem/data is a more natural fit for Time Series algorithms as a starting point.


Hope this helps



Hi John,

Thanks for the quick response.

So, you are saying to first smooth the curve by converting into exponential smoothing and then apply ARIMA ?

how can I do this ?


Can I apply predictor variables in ARIMA in SAS?

Hello Yashu, John and everyone,


I have passion for web analytics problem and will keep an eye on this thread. I am keenly waiting for others' replies too.


My question to you:

Do you account for revisits (Customer coming to the website more than once) to the websites? How did you account for this?


If there are any good references for 'Web analytics', Please share it here.


Many thanks

Magesh N




Just now read a recent "short" article in "Analytics Magazine" titled "Demand Generation: Forecasting traffic for HP online store"... [Analytics Magazine - January/February 2011, check page - 25]

This might be of interest to you as well.


Magesh N


thanks Magesh for sharing


On Data Science Central

© 2020   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service