Mathematical Programming for Missing Data
Chiranjib Bhattacharyya, K.S. Pannagadatta and Alex Smola
In: NIPS, 17, 2004, Dec 2004, Canada.
We propose a mathematical programming method to deal with uncertainty in the
observations of a classification problem. This means that we can deal
with situations where instead of a sample $(\xb_i, y_i)$ we may only
have a distribution over $(\xb_i, y_i)$ at our disposition. In
particular, we derive a robust formulation when the uncertainty is
given by a normal distribution. This leads to Second Order Cone
Programming Problems. Our method can be applied to the problem of
missing data, where it outperforms direct imputation.