ISEN 814 Assignment 2 PDF
Document Details
![GracefulAlien7620](https://quizgecko.com/images/avatars/avatar-20.webp)
Uploaded by GracefulAlien7620
North Carolina A&T State University
Tags
Summary
This document appears to be an assignment for a course on human-machine systems. It includes questions related to signal detection theory and the performance of inspectors in a factory setting, featuring hypothetical scenarios involving detection of flaws in welds.
Full Transcript
ISEN 814 NAME:___________________________________ Advanced Topics in Human Machine Systems Assignment 02 SDT 10 points 0.04 P µ N,σ N 0.02 P µ S_plus_N , σ S_plus_N...
ISEN 814 NAME:___________________________________ Advanced Topics in Human Machine Systems Assignment 02 SDT 10 points 0.04 P µ N,σ N 0.02 P µ S_plus_N , σ S_plus_N Inspector Clouseau 0 20 40 60 80 100 X 0.04 P µ N, σ N 0.02 P µ S_plus_N , σ S_plus_N Inspector Morse 0 20 40 60 80 100 X Inspectors Clouseau and Morse monitor the quality of welds in sheet metal at a factory. The top figure represents the distributions of neural activity (X) of Inspector Clouseau for noise (red solid line) and signal plus noise (blue dashed line). The lower figure represents the activity of Inspector Morse. 1. Which inspector has greater sensitivity (i.e., is better able to detect a flaw in the weld?) (1 pt). 2. Inspector Morse has a neutral criterion, and Inspector Clouseau has a liberal criterion. Draw and label a vertical line representing the criterion in each graph. For which inspector is b < 1? (1 pt). 3. If Inspector Clouseau had a miss rate of.20 and a correct rejection rate of.80, what would his d¢ value be? (Hint: may use attached d’ table, Wickens) (2 pts). 4. The performance of the two inspectors was monitored under different payoff conditions. In the first session they were paid $1 for every detection of a faulty weld, with no penalty for calling a flawless weld faulty. In the second session they were again paid $1 for every correct detection, but were also penalized $1 for every false alarm. In the third session they were paid nothing for correct detections, but were charged $1 for every false alarm. Hit and false alarm rates were computed for each session. Clouseau Morse Hit Rate FA Rate Hit Rate FA Rate Session One.80.44.99.63 Session Two.65.27.85.15 Session Three.45.10.50.02 a) Construct an ROC curve for each observer (plot both curves on the same axes). Label the axes. (2 pts). b) Were both observers affected by the payoffs the same amount, or was one observer more affected by the payoffs? Explain. (1 pt). c) The performance of a third inspector, Inspector Smith, was also monitored under different payoff conditions. His standardized ROC curve, plotted on normal probability paper, had a shallow slope (less than 45 degrees). How would you choose to quantify this inspector’s sensitivity, and why? (1 pt). 5. If Inspector Morse was paid $10 for every hit, but charged $1 for every false alarm, compute the optimal b value for this inspector. (a) Assume P(N)/P(S)=1 (1 pt). (b) Assume that P(N) = 0.8 (1 pt) HINT: Given only P(N) it is possible to determine P(S) and P(N)/P(S)