The idea of StackNet is inpsired by neural network, for learning to rank problems, we have RankNet and ListNet: https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/tr-2007-40.pdf
So is it possible to implement custom objective functions for StackNet so that we can implement a StackNetRanker? I also found that the support for ensembling for learning to rank problems is limited.
The idea of StackNet is inpsired by neural network, for learning to rank problems, we have RankNet and ListNet: https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/tr-2007-40.pdf
So is it possible to implement custom objective functions for StackNet so that we can implement a StackNetRanker? I also found that the support for ensembling for learning to rank problems is limited.