Let φ_X(t) = e^{−t²/2} be the characteristic function of a random variable X. Using the inversion formula, find the probability density function of X.
10 marks · directive: Find
The core of a good answer
Since φ is absolutely integrable, f(x) = (1/2π) ∫ e^{−itx} e^{−t²/2} dt; completing the square gives f(x) = (1/√(2π)) e^{−x²/2}, −∞ < x < ∞, so X is standard normal N(0, 1).
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
The question tests the inversion theorem for characteristic functions: state the condition under which the density formula applies (∫|φ(t)| dt < ∞), then evaluate the Gaussian integral carefully by completing the square, justifying the shift of the integration path.
Answer outline
Introduction
Inversion theorem: if ∫|φ(t)| dt < ∞, X has a continuous density f(x) = (1/2π) ∫_{−∞}^{∞} e^{−itx} φ(t) dt.
Step 1: Check the condition
- ∫ |e^{−t²/2}| dt = √(2π) < ∞, so the density formula applies.
Step 2: Set up
- f(x) = (1/2π) ∫ e^{−itx − t²/2} dt.
Step 3: Complete the square
- −t²/2 − itx = −(t + ix)²/2 − x²/2.
- f(x) = (1/2π) e^{−x²/2} ∫ e^{−(t + ix)²/2} dt.
Step 4: The Gaussian integral
- ∫ e^{−(t + ix)²/2} dt = √(2π) (shift of contour, justified by Cauchy's theorem since the integrand is entire and decays in the strip).
- Alternative: since φ is real and even, f(x) = (1/π) ∫₀^∞ cos(tx) e^{−t²/2} dt = (1/π) √(π/2) e^{−x²/2}.
Step 5: Result
- f(x) = (1/√(2π)) e^{−x²/2}, −∞ < x < ∞.
Conclusion
X ~ N(0, 1). Check: the characteristic function of N(μ, σ²) is e^{iμt − σ²t²/2}, which with μ = 0, σ = 1 is the given φ. The integral was verified symbolically in Python.
Examples and data to cite
- Uniqueness theorem: a characteristic function determines the distribution, so this inversion identifies X completely.
Facts to remember
- Inversion formula: f(x) = (1/2π) ∫ e^{−itx} φ(t) dt when φ is absolutely integrable.
- ∫_{−∞}^{∞} e^{−u²/2} du = √(2π).
- ∫₀^∞ cos(tx) e^{−t²/2} dt = √(π/2) e^{−x²/2}.
- Characteristic function of N(μ, σ²): exp(iμt − σ²t²/2).