Page 240 - AI Ver 3.0 class 10_Flipbook
P. 240
The confusion matrix for the same is:
Prediction
Confusion Matrix
Yes No
Yes 22 47
Actual
No 12 18
Calculate Accuracy, Precision, Recall and F1 Score for the above problem.
(TP+ TN)
Classification Accuracy = × 100
(TP+ TN+FP+FN)
22 18
+
= × 100
( 22 18 ++ 12 + 47)
40
= × 100
99
= 40.40%
TP
Recall =
TP +FN
22
=
22 + 47
22
=
69
= 0.318 or 31.8%
TP
Precision =
TP +FP
22
=
22+12
22
=
34
= .647 or 64.7%
Precision×Recall
F1 Score = 2 ×
Precision+Recall
.64 × .31
= 2 ×
.64 + .31
0.198
= 2 ×
.95
= 0.41
Ethical Concerns Around Model Evaluation
The following are the ethical concerns in the model evaluation:
• Bias - Evaluation metrics may fail to detect biases in a model, leading to unfair outcomes. For example, a
model might favor one gender, race, or socio-economic group over another. To prevent this, metrics should be
carefully designed to avoid introducing or perpetuating bias.
238 Touchpad Artificial Intelligence (Ver. 3.0)-X

