Assignment 5.2 (40 marks) Consider a classifier h that has been evaluated on a test data...

Transcribed Image Text:

Assignment 5.2 (40 marks) Consider a classifier h that has been evaluated on a test data set of n = 100 examples. The observed error is 0.17, i.e. h misclassifies 17 test examples. a) What is the standard deviation and the (two-sided) 95% confidence interval for the true classification error (the error on the entire population)? [Hint: for 95% confidence level, the corresponding value of Z95% is 1.96.] (20 marks) b) How many test examples would you need to ensure that the width of the (two-sided) 95% confidence interval of the true error will be at most 0.1? [Hint: for 95% confidence level, the corresponding value of Z95% is 1.96.] (20 marks) Assignment 5.2 (40 marks) Consider a classifier h that has been evaluated on a test data set of n = 100 examples. The observed error is 0.17, i.e. h misclassifies 17 test examples. a) What is the standard deviation and the (two-sided) 95% confidence interval for the true classification error (the error on the entire population)? [Hint: for 95% confidence level, the corresponding value of Z95% is 1.96.] (20 marks) b) How many test examples would you need to ensure that the width of the (two-sided) 95% confidence interval of the true error will be at most 0.1? [Hint: for 95% confidence level, the corresponding value of Z95% is 1.96.] (20 marks)