RYAN INTERNATIONAL SCHOOL
FARIDABAD
2025-2026
SUBJECT- AI
CLASS- X
TOPIC-AI MODELLING
Ryan OS Assignment
_________________________________________________________________________________________
Learning Outcome: Students will able define Metrics of AI MODELLING
Question: Why should the same data used to train a model not be used for evaluation?
Answer:
Using the same data for both training and evaluation can lead to overfitting, where the model
memorizes the training data and performs well on it but poorly on new, unseen data. To evaluate a
model’s ability to generalize to new data, it is necessary to use a separate test set that was not used
during training. This ensures a more accurate assessment of the model’s performance.
Question: Explain the concept of overfitting in AI models and why it is problematic.
Answer:
Overfitting occurs when a model learns the training data too well, including the noise and outliers,
causing it to perform poorly on new, unseen data. It is problematic because the model fails to
generalize, making incorrect predictions on new data that wasn’t part of the training set. Overfitting
results from excessive complexity in the model, such as too many parameters or insufficient training
data.
Question: What is the confusion matrix, and how does it help in evaluating an AI model?
Answer:
A confusion matrix is a tool used to summarize the performance of a classification model by comparing
the predicted and actual outcomes. It contains four outcomes: True Positive, True Negative, False
Positive, and False Negative. The confusion matrix helps in understanding where the model is making
errors (false positives and false negatives) and gives insight into metrics like accuracy, precision, recall,
and F1 score.
Question: Differentiate between Precision and Recall.
Answer:
Precision measures the proportion of true positive predictions among all positive predictions (True
Positives + False Positives), focusing on the accuracy of positive predictions. Recall measures the
proportion of actual positive cases that were correctly predicted, focusing on how well the model
identifies all actual positive cases (True Positives + False Negatives). Precision is important when the cost
of false positives is high, while Recall is vital when the cost of missing positive cases is significant.
Question: Discuss the importance of F1 Score in model evaluation.
Answer:
The F1 Score is important because it provides a balanced measure of both Precision and Recall, making it
useful when you want to account for both false positives and false negatives. It is especially valuable
when there is an uneven class distribution or when one type of error is more critical than another. The
F1 score ranges from 0 to 1, where 1 represents perfect precision and recall balance.
Question: Why is accuracy not always a reliable metric for model performance, particularly in
imbalanced datasets?
Answer:
In imbalanced datasets, accuracy can be misleading because it may reflect the correct prediction of the
majority class while ignoring the minority class. For example, in a dataset where 95% of cases belong to
one class, a model predicting only the majority class can achieve high accuracy, but it would fail to
predict the minority class correctly, leading to poor performance in real-world applications.
Question: Explain the role of Recall in scenarios where False Negatives are costly.
Answer:
Recall becomes crucial in scenarios where missing a positive case (False Negative) can have severe
consequences. For example, in medical diagnoses or detecting a viral outbreak, failing to identify a
positive case can lead to life-threatening situations. High recall ensures that most of the actual positive
cases are correctly identified, minimizing the occurrence of False Negatives.
Question: How does the F1 Score differ from Accuracy, and when should it be used?
Answer:
The F1 Score differs from Accuracy by focusing on the balance between Precision and Recall rather than
just the overall correct predictions. It should be used in cases where there is an imbalance between
classes or where both False Positives and False Negatives need to be minimized. Accuracy can be
misleading in such cases, while the F1 Score provides a more meaningful evaluation.
Question: What are the four outcomes recorded in a confusion matrix, and what do they signify?
Answer:
The four outcomes in a confusion matrix are:
True Positive (TP): Correctly predicted positive cases.
True Negative (TN): Correctly predicted negative cases.
False Positive (FP): Incorrectly predicted positive cases (Type I error).
False Negative (FN): Incorrectly predicted negative cases (Type II error). These outcomes help assess the
model’s classification performance.
Question: Define Precision and discuss its significance in spam email detection.
Answer:
Precision is the ratio of true positive predictions to the total number of positive predictions (both true
and false positives). In spam email detection, high Precision is important to reduce the number of non-
spam emails incorrectly classified as spam (false positives), as these could contain important
information. A model with high Precision ensures fewer false alarms.
Question: Explain why Recall is important in medical diagnostic models.
Answer:
Recall is vital in medical diagnostics because missing a positive diagnosis (False Negative) can have
serious health consequences. High Recall ensures that most patients with the condition are correctly
identified, minimizing the risk of leaving cases untreated. It reduces the likelihood of False Negatives,
which is critical when early detection can save lives.
Question: What is the role of False Positives and False Negatives in evaluating AI models?
Answer:
False Positives (predicting an event that didn’t happen) and False Negatives (failing to predict an event
that did happen) are crucial in evaluating a model’s performance. False Positives may lead to
unnecessary actions (e.g., sending firefighters when there is no fire), while False Negatives may result in
missing critical events (e.g., ignoring a real fire). Balancing these errors is essential for effective model
evaluation.