Let X follow uniform distribution on (0, 1) and let Y = −logₑ(1 − X). Find the distribution of Y.
5 marks · directive: Find
Answer
The core of a good answer
For y > 0, P(Y ≤ y) = P(X ≤ 1 − e^{−y}) = 1 − e^{−y}, so Y has the exponential distribution with mean 1: f_Y(y) = e^{−y}, y > 0.
An outline written by upsc.help, not an official UPSC answer; every fact checked by a separate review. How we work
What the question asks
Use the distribution-function (CDF) technique or the change-of-variable formula. The transformation is strictly increasing on (0, 1), so the event Y ≤ y can be rewritten directly as an event on X.
Answer outline
Introduction
X ~ U(0, 1), so P(X ≤ u) = u for 0 < u < 1. The map y = −log(1 − x) sends (0, 1) onto (0, ∞) and is strictly increasing.
Step 1: Range
- As x runs from 0 to 1, 1 − x runs from 1 to 0, so y runs from 0 to ∞.
Step 2: CDF
- For y > 0: P(Y ≤ y) = P[−log(1 − X) ≤ y] = P(1 − X ≥ e^{−y}) = P(X ≤ 1 − e^{−y}) = 1 − e^{−y}.
Step 3: Density
- f_Y(y) = d/dy (1 − e^{−y}) = e^{−y}, y > 0; 0 otherwise.
Step 4: Check by Jacobian
- x = 1 − e^{−y}, dx/dy = e^{−y}; f_Y(y) = f_X(x)|dx/dy| = 1 × e^{−y}.
Conclusion
Y ~ exponential with parameter 1 (mean 1, variance 1). This is the basis of generating exponential random numbers from uniform ones.
Examples and data to cite
- Inverse-transform sampling: to simulate an exponential waiting time with mean 1/λ, take −log(1 − U)/λ, U uniform on (0, 1).
Facts
Facts to remember
- CDF technique: F_Y(y) = P[g(X) ≤ y].
- Change of variable: f_Y(y) = f_X(g⁻¹(y)) |d g⁻¹(y)/dy| for monotone g.
- Exponential(λ): F(y) = 1 − e^{−λy}, mean 1/λ.
- If U ~ U(0, 1), then F⁻¹(U) has CDF F (inverse transform).