Fast Nonparametric Conditional Density Estimation

Conditional density estimation. The idea of conditional density estimation is to construct a density estimate f̂(y|x) for a dependent variable y, conditional on a vector of variables x. This can be seen as a generalization of regression, where instead of estimating the expected value E(y|x) alone, we instead model the full density. This is especially important for multi-modal densities, where the expected value might be nowhere near a mode, and for situations in which confidence intervals are preferred to point estimates. Some problems that can be addressed by conditional density estimates are: time series prediction, static regression with confidence bands, learning continuous k-Markov models, and collaborative filtering.