The table below shows a list of 20 observations in a sample data set and an estimated probability of Class 1 membership given by the data mining software. Assume that the cutoff value is 0.50 .
able[[Actual Class, able[[Probability of Class],[1]],Actual Class, able[[Probability of Class],[1]]],[1,1.00,0,0.70],[0,1.00,0,0.68],[1,0.98,1,0.47],[1,0.95,0,0.46],[0,0.92,0,0.45],[0,0.88,0,0.44],[1,0.85,0,0.42],[0,0.78,0,0.35],[0,0.77,1,0.22],[1,0.72,0,0.08]]
Based on the given information, a classification confusion matrix can be constructed as below:
able[[,Predicted Class],[Actual Class,0,1],[0,n_(00)=dots-,n_(01)=dots
Question 44(3 points)
In general,if we increase the cutoff value.Class1 error rate
will stay constant
will decrease
will increase
may increase or decrease,depending on the sample data