gppe

Gaussian Process Preference Elicitation (GPPE) package. Implements the preference elicitation model of Bonilla et al [1].

Author:

Edwin V. Bonilla Last update: 22/05/2012

Main Contents

1. elicit_gppe.m: Elicit preferences for a new user with the GPPE model.
2. learn_gppe.m: Computes the negative marginal likelihood and its gradients wrt to the hyper-paramters. This can be used to learn a gppe model with Carl Rasmussen's minimize function
3. toy_example.m: A toy example demonstrating the use of the gppe package

References

[1] Edwin V. Bonilla, Shengbo Guo and Scott Sanner. Gaussian process preference elicitation. In Advances in Neural Information Processing Systems 23: NIPS'10.