Simulation DATA-SCIENCE-OPTIMIZE Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A financial analyst is using a Monte Carlo simulation to estimate the value at risk (VaR) of a portfolio. The simulation uses 10,000 iterations. The analyst is concerned about the computational cost and wants to reduce the number of iterations while maintaining accuracy. Which approach is most appropriate?

   • AUse variance reduction techniques such as antithetic variates or control variates to reduce the variance of the estimate, allowing fewer iterations for the same accuracy.
   • BSimply reduce the number of iterations to 1,000 and accept the lower accuracy.
   • CUse a deterministic approximation of the portfolio returns instead of simulation.
   • DUse bootstrapping on the historical returns to generate new samples.
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   How can simulation be used to validate a predictive model?

   • ABy removing all randomness from the model to make it deterministic
   • BBy generating synthetic data from known distributions to test model behavior
   • CBy increasing the sample size of the original dataset
   • DBy using the model to predict future real-world events with certainty
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   What is the primary purpose of Monte Carlo simulation in data science?

   • ATo generate a single deterministic prediction from a model
   • BTo replace the need for real-world data collection entirely
   • CTo optimize a model by adjusting its parameters manually
   • DTo model the impact of uncertainty by running many trials with random inputs
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A supply chain analyst is simulating the total lead time for an order, which is the sum of supplier processing time (uniformly distributed between 2 and 5 days) and shipping time (normally distributed with mean 3 days and standard deviation 1 day). The analyst needs to estimate the probability that total lead time exceeds 8 days. Which simulation approach is correct?

   • ADraw a single processing time from the uniform distribution and a single shipping time from the normal distribution, sum them, and use that as the estimate.
   • BUse a normal approximation for the total with mean 6.5 and standard deviation 1, then find the z-score for 8.
   • CUse the sum of the means (3.5 + 3 = 6.5) and conclude the probability is low because 8 is far above the mean.
   • DFor each iteration, draw a processing time from the uniform distribution and a shipping time from the normal distribution, sum them, and record the total. After many iterations, compute the proportion of totals above 8 days.
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   A data scientist is comparing the performance of two machine learning models on a small dataset (n=50). The scientist wants to determine if the difference in accuracy between the models is statistically significant. Which approach is most appropriate?

   • AUse cross-validation to get multiple accuracy scores for each model and then perform a t-test on the differences.
   • BUse a chi-squared test on the confusion matrices of the two models.
   • CUse a paired bootstrap test: resample the dataset with replacement, evaluate both models on each resample, compute the difference in accuracy, and compare the observed difference to the distribution of differences.
   • DUse a standard t-test on the accuracy scores from a single train-test split.
   Reveal Solution Discussion Select an answer first