Question

A good statistic to use to evaluate imbalanced classification problems is?

		overall accuracy
		Matthews correlation coefficient
		F1
		sensitivity

A credit risk modeler wants to ensure the propensity to default scores he generated for each customer is as accurate as possible. Which approach might help here?

		gains chart
		probability calibration plot
		overall accuracy
		lift chart

A classification problem realized TP=10; TN=9; FP=15; FN=14. Which statistic is the greatest?

		sensitivity
		NPV
		specificity
		PPV

Which of the following is not a regression problem statistical performance measure?

		MAE
		MSE
		log-loss
		sMAPE

Answer 1

(Q-1)A good statistic to use to evaluate imbalanced classification problems is F1.

Answer: F1

Hence the "option-3" is the correct answer.

(Q-2)A credit risk modeler wants to ensure the propensity to default scores he generated for each customer is as accurate as possible is probability calibration plot.

Answer: probability calibration plot.

Hence the "option-2" is the correct answer.

(Q-3)A classification problem realized TP=10; TN=9; FP=15; FN=14. specificity statistic is the greatest.

Answer: specificity

Hence the "option-3" is the correct answer.

(Q-4)Which of the following is not a regression problem statistical performance measure?

Answer: sMAPE is not a regression problem statistical performance measure.

Hence the "option-4" is the correct answer.