| There are lot of situations where heavy-tailed models have proven to be a good fit, and at the same time there is a natural upper bound on the possible values, and hence the need to study truncated heavy-tailed models. This talk is on understanding such models. It turns out that depending upon on the truncation level and the tail, one can differntiate between two regimes. To be specific, the difference between the two regimes shows up in the central limit behaviour and the decay rate of large deviation prbabilities. The last part of the talk is on the statistical problem of deciding the truncation regime from data.
This is a joint work with Prof. Gennady Samorodnitsky. |