Probability theory MOC

Jensen’s inequality

Let (πœ‰,F,β„™) be a probability model, 𝑋 :πœ‰ →ℝ be a real random variable1, and πœ‘ :ℝ →ℝ be a Convex function. Then measure

πœ‘(βˆ«πœ‰π‘‹π‘‘β„™)β‰€βˆ«πœ‰πœ‘βˆ˜π‘‹π‘‘β„™πœ‘(𝔼⁑[𝑋])≀𝔼⁑[πœ‘(𝑋)]

with equality iff πœ‘(𝑋) =π‘Ž +𝑏𝑋.


develop | en | SemBr

Footnotes

  1. i.e. β„™-measurable function ↩