Achieves 0.55 MRR (and 0.64 MRR on the specific subset of data) at the RecSys2019 challenge using a two stage deep neural network architecture. More information can be gathered from the Documentation.pdf in particular at the "Deep Learning Model" section..
Autoencoder für die Produkte implementieren, der nicht nur als Ziel hat das Items zu rekonstruieren, sondern auch möglichst unterschiedlich zu anderen Items in der selben Session zu sein
Autoencoder für die Produkte implementieren, der nicht nur als Ziel hat das Items zu rekonstruieren, sondern auch möglichst unterschiedlich zu anderen Items in der selben Session zu sein