next basket recommendation about sequence-based-recommendations

[rdevooght]

Hi,

I am not fully versed in the specificities of next-basket recommendation, but I think that there is certainly a way to adapt the RNN approach to that problem.
In the following paper: http://bit.ly/umap_paper we compare the RNN with FPMC, which was originally designed for next-basket recommendation. In those experiments FPMC was simplified to deal with the standard item prediction problem, but you could take the other direction and adapt the RNN to the next-basket case.
Some have done that for feedforward neural network: http://ceur-ws.org/Vol-1441/recsys2015_poster15.pdf
It can be an inspiration for what you would have to do.

Best of luck,
Robin

[richielo]

Hi Robin,

Thank you very much for the reply. i will definitely look into it.

Richie