1-3 Monte Carlo runs for algorithm comparison. Two estimation algorithms are to

Question

1-3 Monte Carlo runs for algorithm comparison. Two estimation algorithms are to be compared based on Monte Carlo runs. Assume their average performances (MSE) were obtained as C(1) = 5 and C(2) = 8 from N = 100 runs. 1. Can one say that algorithm 1 is superior to algorithm 2? 2. Assume the standard deviations associated with these sample means are 1/2 )-Cu) Using the above information answer question 1. 3. Denote the correlation coefficient between the outcomes of the runs with the two algorithms as p, that is, 0.5 and the variances are as in item 2, the optimal statistical test will Show that if = yield a conclusive answer

Explanation / Answer

First of all One can conclude that Algorithm 1 is better or Algorithm 2 is better only based on average performance. Only averages are not sufficient to conclude. It needs additional information on variability. Give SD for Algorithm to be same and is 1, it means that CV for Algorithm 1 is Mean/SD = 5 and for Algorithm 2 it is 8. Therefore based on CV we can conclude that Algorithm 1 is more consistent as compared to other. Under given condition stated expectation is 0.5 only. It means that chance of one Algorithm is superior over other is 50 %. Means both Algorithms may be treated as either better or worst simultaneously.

Related Questions

Navigate

• Browse (All)
• Browse 1
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.