Efficient Statistics, in High Dimensions, from Truncated Samples
2018 IEEE 59TH ANNUAL SYMPOSIUM ON FOUNDATIONS OF COMPUTER SCIENCE (FOCS), pp. 639.0-649, 2018.
efficient statisticstruncated samplesnormal distributionhigh dimension
We provide an efficient algorithm for the classical problem, going back to Galton, Pearson, and Fisher, of estimating, with arbitrary accuracy the parameters of a multivariate normal distribution from truncated samples. Truncated samples from a d-variate normal means a samples is only revealed if it falls in S of the d-dimensional Euclide...More
Full Text (Upload PDF)
PPT (Upload PPT)