GATE DA Statistical Computing: Simulation, Resampling & Algorithms: practice questions
GATE - Data Science and Artificial Intelligence (DA) · India · IITs and IISc on rotation (2027: IIT Madras)
Original practice questions written to the published syllabus · not past papers
10worked here, free
20written to this exam in this subject
11syllabus units
About GATE DA
GATE - Data Science and Artificial Intelligence (DA) is a postgraduate admission exam run by IITs and IISc on rotation (2027: IIT Madras), for candidates in India. Computer-based, 3 hours, 100 marks (15 General Aptitude + 85 subject). MCQ, MSQ and NAT; negative marking on MCQ only.
Checked against the exam body's published information. Formats and fees change; confirm the current edition there before you book.
Statistical Computing: Simulation, Resampling & Algorithms in GATE DA
The algorithms inside modern statistics: random-number generation, Monte Carlo integration, the bootstrap and jackknife, permutation tests, the EM algorithm, MCMC and numerical linear algebra, with attention to accuracy and reproducibility.
Our 20 GATE DA questions in this subject: 10 easy, 10 medium, 0 hard, across 11 units.
10 practice questions, worked
Choose an answer, then open the working. Each explanation says why the right answer is right and why the others are not.
These 10 were written to GATE DA and are not shown on any other exam's page.
Question 1 · easy
For the generator xₙ₊₁ = (5xₙ + 3) mod 16 with x₀ = 7, what is x₁?
38
6
2
7
Answer and working
B. 6
5 × 7 + 3 = 38, and 38 mod 16 = 6; 38 forgets the modulus.
Question 2 · medium
Setting a random seed before a simulation mainly ensures that:
The numbers become truly random in every run
The simulation runs faster
No two runs can ever agree
Results can be reproduced
Answer and working
D. Results can be reproduced
Pseudo-random sequences are deterministic given the seed.
Question 3 · medium
The inversion method generates X from distribution F as:
F⁻¹(U), with U uniform on (0, 1)
F(U), with U uniform on (0, 1)
U/F(U)
1 − F(U)²
Answer and working
A. F⁻¹(U), with U uniform on (0, 1)
The probability integral transform runs in reverse.
Question 4 · easy
To generate a Bernoulli(p) variable from U uniform on (0, 1), return 1 if:
U > p²
U = p
U < p
U < 1 − p²
Answer and working
C. U < p
P(U < p) = p for a uniform U, so the result is 1 with probability p.
Question 5 · easy
The standard error of a Monte Carlo estimate from n independent draws decreases in proportion to:
1/√n
1/n
1/n²
log n
Answer and working
A. 1/√n
Halving the error needs four times as many draws.
Question 6 · medium
A control variate with correlation 0.9 to the integrand, used optimally, reduces variance by about:
81%
90%
10%
19%
Answer and working
A. 81%
Variance is multiplied by 1 − ρ² = 0.19.
Question 7 · medium
A 95% percentile bootstrap interval uses:
θ̂ ± 1.96 times the data SD
The minimum and maximum of θ̂*
The 5th and 95th percentiles
The 2.5th and 97.5th percentiles of θ̂*
Answer and working
D. The 2.5th and 97.5th percentiles of θ̂*
The normal-theory interval would use the bootstrap SE instead.
Question 8 · easy
The jackknife also supplies, in BCa bootstrap intervals, the:
Number of resamples
Seed
Confidence level
Acceleration constant
Answer and working
D. Acceleration constant
The acceleration is estimated from jackknife influence values.
Question 9 · easy
How many distinct ways are there to assign 10 observations into two groups of 5?
100
3,628,800
32
252
Answer and working
D. 252
C(10, 5) = 252; 3,628,800 is 10!.
Question 10 · easy
Bootstrap resamples and simulation replicates are called 'embarrassingly parallel' because they:
Tasks are independent
They cannot be parallelised on any hardware
Must run in strict sequence
Need one shared random number
Answer and working
A. Tasks are independent
Each worker can run its own share.
A study plan for Statistical Computing: Simulation, Resampling & Algorithms in GATE DA
This plan follows where our questions actually fall, so the time goes where the marks are likely to be. It is our reading of the syllabus, not the exam body's weighting.
First, the heavy units (45% of the 20 questions): Pseudo-random number generation and tests of randomness (3, 15%); Bootstrap: percentile and BCa intervals (2, 10%); Generating random variates: inversion, acceptance-rejection and transformation methods (2, 10%); Jackknife (2, 10%).
Read the manual chapters for these first and do every worked question you can find on them.
Then the middle band: MCMC algorithms in practice (2, 10%); Monte Carlo integration and variance reduction (2, 10%); Parallel and high-performance computing for statistics (2, 10%); Permutation and randomisation tests (2, 10%).
Aim for one sitting per unit, then mixed questions across them.
Then the lighter units: EM algorithm and its applications (missing data, mixtures) (1, 5%); Numerical linear algebra for statistics: QR and Cholesky (1, 5%); Simulation studies: design and reporting (1, 5%).
Learn the definitions, classifications and standard numbers; these units are more often recalled than worked.
Bank the easy marks: Jackknife (2 easy of 2); Pseudo-random number generation and tests of randomness (2 easy of 3).
Do these early in revision so they are fast on the day.
Finish timed. Sit the subject as a timed set without notes, mark it, and send each wrong answer back to its unit above.