Hi, I'm looking for a way to define an algorithm to predict marketing campaigns impacts, based on historic data.

I was thinking about time series models but don't know if they're the best approach

Thanks

If we assume that the marketing campaign are always similar, then time series could be ok because it would only considers the number and ignore other aspects.

But marketing campaigns are probably different each time and the other aspects should perhaps be taken into account like the time of the year, and perhaps many others factors related to how the marketing is done or about the customers, the products (just for example).

An idea could be to try a neural network where the input is all the relevant factors and the output is some results (good sales, bad sales, etc.).

I cannot tell much more since I did not work on this problem and I did not read about this. But that is my quick opinion about this.

Great I'll try with neural network, thanks!

i need data mining topics for dissertation