Skip to content

FRM Part I · FRM Exam Part I · Machine-Learning Methods

Using the confusion matrix of 30 TP, 10 FP, 20 FN and 140 TN from a default classifier at a 0.5 threshold, the bank lowers the cutoff to 0.3 and the new counts are 40 TP, 30 FP, 10 FN and 120 TN. Which statement correctly describes the change?

Accuracy falls from 0.85 to 0.80, recall rises from 0.60 to 0.80, and precision falls from 0.75 to 0.571. Lowering the threshold flags more loans as defaults, catching more true defaults but also producing more false alarms.

  1. AAccuracy rises from 0.85 to 0.80 and the F1 score rises from 0.667 to 0.667
  2. BAccuracy falls from 0.85 to 0.80 and the F1 score falls from 0.667 to 0.571
  3. CAccuracy rises from 0.85 to 0.80 and recall falls from 0.60 to 0.57
  4. DAccuracy falls from 0.85 to 0.80 while recall rises from 0.60 to 0.80 and precision falls from 0.75 to 0.571Correct

Explanation

Before: accuracy = 170/200 = 0.85, recall 0.60, precision 0.75. After: accuracy = 160/200 = 0.80, recall = 40/50 = 0.80, precision = 40/70 = 0.571. Lowering the cutoff trades precision for recall. Other options misstate the direction or values.

Did you get it right without looking?

One question tells you little. A timed set on Machine-Learning Methods shows your real accuracy, how long you take and where you lose marks.

Practise Machine-Learning MethodsFree diagnostic mock

More Machine-Learning Methods questions