Page 246 - AI Ver 3.0 class 10_Flipbook
P. 246
The term Error means the action that is inaccurate or wrong. It refers to the difference between a model’s prediction
and the actual outcome. It quantifies how often the model makes mistakes. Based on the error, we choose the machine
learning model that has the best performance for a specific dataset. Low Error in a model performance signifies precise
and reliable predictions.
5. Discuss the ethical concerns around model evaluation.
Ans. 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.
• Fairness: Fairness ensures that models treat all groups equally. The evaluation processes must account for fairness
to prevent models from producing discriminatory results.
• Transparency: Sometimes, evaluation models lack clarity, making it difficult to understand how predictions are made.
A transparent approach clearly explains how metrics are chosen and how results are derived.
• Accountability: It is crucial to take responsibility for the choice of evaluation metrics and their outcomes. Providing
clear reasoning behind metric selection helps ensure ethical decision-making and builds trust.
• Privacy: Using real-world data for evaluation often involves sensitive personal information, raising concerns about
privacy. Measures should be in place to ensure that individual data is protected during the evaluation process.
• Data Protection: Protecting the data used in model evaluation is essential to prevent misuse or unauthorised access.
C. Competency-based/Application-based questions. 21 st Century #Critical Thinking
Skills
1. An AI model made the following predictions for Book Sales forecast. Calculate Accuracy, precision and recall for the
following confusion matrix:
Confusion Matrix True Positives True Negatives
Predicted Positive 50 40
Predicted Negative 12 10
Correction prediction
Ans. Accuracy = × 100%
Total Cases
(TP + TN)
= × 100%
(TP + TN + FP + FN)
50 + 10
= × 100%
50 + 10 + 40 + 12
60
= × 100%
112
= 53.5%
True Positive
Precision =
All Predicted Positive
TP
=
TP + FP
50
=
50 + 40
= 0.555 or 55.5%
244 Touchpad Artificial Intelligence (Ver. 3.0)-X

