AUC |
0.779361 |
0.500000 |
Area under the ROC curve (true pos. rate vs false pos. rate); ranges from 0.5 (purely random classifier) to 1.0 (perfect classifier) |
True Positive Rate |
0.338983 |
0.000000 |
33.9% of positive instances were correctly identified.; i.e. 20 "Defaulted" labels were correctly identified out of 59 instances; a.k.a Sensitivity/Recall |
True Negative Rate |
0.936170 |
1.000000 |
93.6% of negative instances were correctly identified.; i.e. 132 "Not Defaulted" labels were correctly identified out of 141 instances |
False Positive Rate |
0.063830 |
0.000000 |
6.4% of negative instances were incorrectly identified as positive; i.e. 9 "Not Defaulted" labels were incorrectly identified as "Defaulted", out of 141 instances |
False Negative Rate |
0.661017 |
1.000000 |
66.1% of positive instances were incorrectly identified as negative; i.e. 39 "Defaulted" labels were incorrectly identified as "Not Defaulted", out of 59 instances |
Positive Predictive Value |
0.689655 |
0.000000 |
When the model claims an instance is positive, it is correct 69.0% of the time; i.e. out of the 29 times the model predicted "Defaulted", it was correct 20 times; a.k.a precision |
Negative Predictive Value |
0.771930 |
0.705000 |
When the model claims an instance is negative, it is correct 77.2% of the time; i.e. out of the 171 times the model predicted "Not Defaulted", it was correct 132 times |
F1 Score |
0.454545 |
0.000000 |
The F1 score can be interpreted as a weighted average of the precision and recall, where an F1 score reaches its best value at 1 and worst score at 0. |
Precision/Recall AUC |
0.588795 |
0.295000 |
Precision/Recall AUC is calculated with `average_precision` which summarizes a precision-recall curve as the weighted mean of precisions achieved at each threshold. See sci-kit learn documentation for caveats. |
Accuracy |
0.760000 |
0.705000 |
76.0% of instances were correctly identified |
Error Rate |
0.240000 |
0.295000 |
24.0% of instances were incorrectly identified |
% Positive |
0.295000 |
0.295000 |
29.5% of the data are positive; i.e. out of 200 total observations; 59 are labeled as "Defaulted" |
Total Observations |
200 |
200 |
There are 200 total observations; i.e. sample size |