Discussion Overview
The discussion revolves around the expectation of the logarithm of a random variable, specifically whether Elog(x) can be infinite for certain distributions. The scope includes theoretical considerations and implications of Jensen's inequality in the context of probability and statistics.
Discussion Character
- Exploratory
- Technical explanation
- Debate/contested
Main Points Raised
- One participant suggests that by Jensen's inequality, Elog(x) is less than or equal to logE(x), which is finite if the mean is finite.
- Another participant asserts that since x > 0, log(x) must always be finite, implying E(log(x)) must also be finite.
- However, a different viewpoint is presented that as x approaches 0, log(x) approaches -infinity, raising the possibility of Elog(x) being -infinity.
- A further suggestion is made to consider the distribution x = exp(-1/u) where u is uniform on (0,1) as a potential counterexample.
Areas of Agreement / Disagreement
Participants express differing views on whether Elog(x) can be -infinity, indicating that the discussion remains unresolved with competing perspectives on the implications of the logarithmic behavior of x.
Contextual Notes
There are limitations regarding the assumptions about the distribution of x and the behavior of log(x) as x approaches 0, which are not fully explored in the discussion.