One mark Questions

Fill in the Blanks

1. Techniques such as descriptive statistics and visualization can be applied to the

data set, to assess the content, quality, and initial insights about the data
2. Selecting the AI or machine learning use cases and being able to evaluate the return
on investment (ROI) is critical to the success of any data project.
3. Train Dataset is used to fit the machine learning model.
4. Loss functions can be broadly categorized into Classification and Regression Loss
5. A good model should have an RMSE value <180
6. Regulatory compliance testing and security testing are important since the system
might deal with sensitive data.
7. Human biases in selecting test data can adversely impact the testing phase, therefore
data validation is important.
8. After clearly stating a business problem, the data scientist can define the analytic
approach to solve it.
9. In the analytic approach in order to o show relationships, descriptive approach
may be required.
10. The data scientist will use a training set for predictive modelling.
11. Test Dataset is used to evaluate the fit machine learning model.
12. Mean Square Error (MSE) is the most commonly used regression loss function.
13. The first fundamental step when starting an AI initiative is scoping.
14. In the analytic approach in order to determine probabilities of an action, then
predictive model might be used.
15. Train-test split is a technique for evaluating the performance of a machine learning
algorithm.
16. Cross-validation gives you a more reliable measure of your model's quality, though
it takes longer to run.
17. Performance metrics like classification accuracy and root mean squared error can
give you a clear objective idea of how good a set of predictions.
18. A loss function is a measure of how good a prediction model does in terms of
being able to predict the expected outcome.
19. In order to find the minimum point of function gradient descent is used.
20. The final project of an academic program called Capstone Project
21. Train-Test Split is a common AI model validation technique.
22. Anomaly Detection is an AI technique is used to detect unusual patterns.
23. Trend is the term that describes increasing or decreasing values in time series data.

Multiple Choice Questions

1. What is the first step in the AI Project Cycle?

A) AI Model Construction
B) Data Gathering
C) Problem Definition
D) Deployment
Answer: ✅ C) Problem Definition

2. Which of the following is NOT a key stage of the AI model life cycle?
A) Data Collection
B) Model Training
C) Deployment
D) Internet Browsing
Answer: ✅ D) Internet Browsing

3. What does RMSE stand for in AI model evaluation?

A) Root Mean Squared Error
B) Random Model Selection Error
C) Recurrent Machine Standard Evaluation
D) Regression Mean Standard Error
Answer: ✅ A) Root Mean Squared Error

4. Which technique is used to classify data into predefined categories?

A) Regression
B) Clustering
C) Classification
D) Anomaly Detection
Answer: ✅ C) Classification

5. In cross-validation, why is the dataset split into multiple subsets?

A) To improve model performance and reduce overfitting
B) To reduce computation time
C) To eliminate the need for training
D) To test only a small part of the dataset
Answer: ✅ A) To improve model performance and reduce overfitting

6. What is the purpose of feature engineering in AI?

A) To increase model accuracy by modifying input variables
B) To create complex AI algorithms
C) To deploy the AI model in production
D) To store large datasets
Answer: ✅ A) To increase model accuracy by modifying input variables

7. What is the main advantage of using a train-test split in AI?

A) It ensures the model works well on unseen data
B) It eliminates the need for feature selection
C) It allows models to memorize the training data
D) It reduces the size of the dataset
Answer: ✅ A) It ensures the model works well on unseen data

8. Which AI technique is best suited for detecting fraudulent transactions?

A) Classification
B) Regression
C) Anomaly Detection
D) Clustering
Answer: ✅ C) Anomaly Detection

9. What is the primary goal of Data Storytelling?

A) To make data insights engaging and easy to understand
B) To increase the size of the dataset
C) To replace traditional statistics
D) To eliminate the need for data visualization
Answer: ✅ A) To make data insights engaging and easy to understand

10. Which AI model validation technique involves splitting data into K subsets
and testing on each subset?
A) Train-Test Split
B) Decision Trees
C) K-Fold Cross-Validation
D) Feature Engineering
Answer: ✅ C) K-Fold Cross-Validation

11. What is the main purpose of Design Thinking in AI?

A) To design user interfaces
B) To create AI models without data
C) To develop human-centered AI solutions
D) To replace machine learning algorithms
Answer: ✅ C) To develop human-centered AI solutions

12. Which AI model evaluation metric is most sensitive to outliers?

A) RMSE
B) MSE
C) Mean Absolute Error (MAE)
D) Accuracy
Answer: ✅ B) MSE (Mean Squared Error)

13. Which of the following is NOT an example of AI-based predictive analysis?

A) Stock price prediction
B) Spam email detection
C) Weather forecasting
D) Image compression
Answer: ✅ D) Image compression

14. What is the role of data visualization in AI storytelling?

A) To remove irrelevant data
B) To make patterns in data more understandable
C) To replace AI models
D) To create raw data
Answer: ✅ B) To make patterns in data more understandable

15. In machine learning, what does overfitting mean?

A) The model performs well on training data but poorly on new data
B) The model has too many missing values
C) The model is too simple and cannot learn patterns
D) The model does not use training data
Answer: ✅ A) The model performs well on training data but poorly on new data
[Link] is the main objective of a Capstone Project in AI?
A) To memorize AI concepts
B) To apply AI knowledge to real-world problems
C) To create large datasets
D) To replace traditional programming
✅ Answer: B) To apply AI knowledge to real-world problems

17. Which of the following is NOT a phase of the AI Project Cycle?

A) Problem Definition
B) Feature Selection
C) Social Media Marketing
D) Model Evaluation
✅ Answer: C) Social Media Marketing

18. What is the first step in problem decomposition?

A) Writing the code
B) Testing the model
C) Understanding and restating the problem
D) Choosing a programming language
✅ Answer: C) Understanding and restating the problem

19. Which AI method is used to group similar data points together?

A) Regression
B) Classification
C) Clustering
D) Anomaly Detection
✅ Answer: C) Clustering

20. In AI problem-solving, what is the importance of breaking down a problem

into smaller parts?
A) It makes coding easier
B) It allows parallel processing
C) It helps in structured problem-solving
D) It eliminates the need for data
✅ Answer: C) It helps in structured problem-solving
Model Lifecycle & AI Model Validation
21. What is the purpose of the AI Model Lifecycle?
A) To ensure AI models are built and maintained efficiently
B) To eliminate the need for model evaluation
C) To store large datasets
D) To create random AI models
✅ Answer: A) To ensure AI models are built and maintained efficiently

22. Which of the following is a commonly used AI validation technique?

A) Train-Test Split
B) Waterfall Model
C) Data Merging
D) Feature Deletion
✅ Answer: A) Train-Test Split

23. Which of the following is NOT a key concept in AI Model Validation?

A) RMSE
B) MSE
C) Mean Absolute Error
D) Video Compression
✅ Answer: D) Video Compression

24. Why is cross-validation used in machine learning?

A) To evaluate model performance on different data subsets
B) To reduce dataset size
C) To remove noisy data
D) To replace neural networks
✅ Answer: A) To evaluate model performance on different data subsets

25. What is the main advantage of K-Fold Cross-Validation over Train-Test

Split?
A) It requires no data preprocessing
B) It provides a more reliable estimate of model performance
C) It reduces computing costs
D) It eliminates the need for data collection
✅ Answer: B) It provides a more reliable estimate of model performance

AI Techniques & Algorithms

26. Which machine learning technique is used for predicting continuous
values?
A) Classification
B) Regression
C) Clustering
D) Reinforcement Learning
✅ Answer: B) Regression

27. What is the role of Feature Engineering in AI?

A) To create new input variables that improve model accuracy
B) To store large datasets
C) To reduce data storage needs
D) To replace AI algorithms
✅ Answer: A) To create new input variables that improve model accuracy

28. What is the purpose of an AI recommendation system?

A) To suggest relevant items based on user behavior
B) To classify spam emails
C) To detect fraudulent transactions
D) To analyze customer sentiment
✅ Answer: A) To suggest relevant items based on user behavior

Data Science & Storytelling Through Data

29. What is the primary goal of storytelling in AI?
A) To make data insights clear and compelling
B) To generate random numbers
C) To eliminate the need for data analysis
D) To replace predictive models
✅ Answer: A) To make data insights clear and compelling

30. Which of the following is an important aspect of data storytelling?

A) Using only raw data
B) Ignoring audience understanding
C) Combining visuals, data, and narrative
D) Replacing statistical analysis
✅ Answer: C) Combining visuals, data, and narrative

31. How does visualization enhance data storytelling?

A) By making patterns easier to interpret
B) By replacing mathematical models
C) By increasing data storage
D) By reducing dataset size
✅ Answer: A) By making patterns easier to interpret

32. What are the key elements of a good data story?

A) Context, visualization, and narrative
B) Only numerical data
C) Raw data without interpretation
D) Unstructured text analysis
✅ Answer: A) Context, visualization, and narrative
33. Why are AI-powered chatbots effective in customer service?
A) They eliminate the need for human support
B) They provide automated responses 24/7
C) They store large amounts of data
D) They replace human intelligence
✅ Answer: B) They provide automated responses 24/7

AI Evaluation Metrics
34. What does Mean Squared Error (MSE) measure?
A) The squared differences between actual and predicted values
B) The sum of all predicted values
C) The average classification accuracy
D) The number of training iterations
✅ Answer: A) The squared differences between actual and predicted values

35. What is a limitation of Mean Absolute Error (MAE)?

A) It does not penalize large errors enough
B) It is difficult to compute
C) It is not used in regression problems
D) It ignores all errors
✅ Answer: A) It does not penalize large errors enough

Short Answer Type Questions

1. Define a Capstone Project.
o A Capstone Project is a final academic project where students apply
their learning to solve real-world problems and demonstrate their
knowledge.
2. What are the key steps in an AI project cycle?
o Problem Definition, Data Gathering, Feature Definition, AI Model
Construction, Evaluation & Refinements, Deployment.
3. Differentiate between Classification and Regression.
o Classification categorizes data into predefined classes, whereas
Regression predicts continuous numerical values.
4. What are the five stages of Design Thinking?
o Empathize, Define, Ideate, Prototype, and Test.
5. What is RMSE, and why is it used?
o Root Mean Squared Error measures the difference between actual and
predicted values in regression models, helping assess accuracy.
6. Explain the importance of Data Validation in AI.
o Data Validation ensures the correctness, consistency, and reliability of
data before model training to improve accuracy and avoid bias.
7. What are the three main stages of an AI model life cycle?
o Project Scoping, Model Building, and Deployment.
8. Why is storytelling important in data science?
 o Storytelling enhances data interpretation, making insights more
compelling and understandable for decision-making.
9. What is Cross-Validation in AI model evaluation?
o Cross-Validation is a technique where a dataset is split into multiple
subsets to improve model accuracy and reduce bias.
10. How does Anomaly Detection help in AI applications?
o It helps identify outliers or unusual patterns in data, useful for fraud
detection, medical diagnosis, and security monitoring.

Capstone Project & AI Project Cycle

1. What is the purpose of a Capstone Project in AI?
A Capstone Project is designed to integrate all the AI concepts learned throughout the
course and apply them to a real-world problem. It allows students to demonstrate their
ability to solve complex problems using AI techniques, including data collection,
model building, and evaluation.
2. List any three key concepts used in AI Capstone Projects.
1. AI Project Cycle – The structured approach for developing AI projects.
2. Model Validation – Ensuring the model is performing correctly by testing
it on different datasets.
3. RMSE (Root Mean Squared Error) – A metric to measure the accuracy
of regression models.
3. What are the six steps of the AI Project Cycle?
1. Problem Definition – Understanding and defining the problem.
2. Data Gathering – Collecting relevant data.
3. Feature Definition – Selecting important attributes of the data.
4. AI Model Construction – Training the AI model.
5. Evaluation & Refinements – Testing and improving the model.
6. Deployment – Implementing the model in a real-world scenario.
4. Why is it important to define the problem before applying AI techniques?
Defining the problem ensures that AI is used appropriately and helps in selecting the
right approach and model. If the problem is not well understood, applying AI
techniques may lead to incorrect or ineffective solutions.
5. What are the five main types of questions that predictive analysis answers?
1. Which category? (Classification) – E.g., Is this email spam or not?
2. How much or how many? (Regression) – E.g., What will be the price of a
stock tomorrow?
3. Which group? (Clustering) – E.g., Customer segmentation in marketing.
4. Is this unusual? (Anomaly Detection) – E.g., Fraud detection in banking.
5. Which option should be taken? (Recommendation) – E.g., Personalized
product recommendations on Amazon.
6. Give an example of how AI can help in the agriculture sector.
AI can be used in agriculture for crop prediction by analyzing weather patterns, soil
conditions, and past yield data. For example, AI models can predict the best time to
plant crops, detect diseases early, and optimize irrigation schedules.
7. What is the role of Design Thinking in AI problem-solving?
Design Thinking is a structured approach to problem-solving that focuses on
understanding the user’s needs before developing an AI solution. It helps in creating
AI models that are more user-friendly, effective, and aligned with real-world
requirements.
8. List the five stages of Design Thinking.
1. Empathize – Understand the user’s problems and challenges.
2. Define – Clearly define the problem that needs to be solved.
3. Ideate – Brainstorm multiple AI solutions.
4. Prototype – Develop a small-scale version of the AI model.
5. Test – Validate and refine the solution based on real-world feedback.
9. Why is problem decomposition important in AI projects?
Problem decomposition helps break complex AI tasks into smaller, manageable
sub-tasks. This improves the efficiency of AI solutions, reduces errors, and simplifies
debugging.
10. How does breaking down a complex problem into smaller steps improve AI
solutions?
By breaking a problem into smaller steps, each part can be tackled separately, making
the AI model more efficient, accurate, and scalable. For example, in a movie
recommendation system, individual tasks like user data analysis, content filtering,
and trend detection are handled separately.

AI Model Life Cycle & AI Model Validation

11. What are the three main stages of the AI Model Life Cycle?
1. Project Scoping – Defining business objectives, identifying AI use cases,
and planning the project.
2. Model Building – Collecting data, selecting the right algorithm, and
training the model.
3. Deployment & Monitoring – Implementing the model in production and
continuously improving it.
12. Why is data validation important in AI model building?
Data validation ensures that the input data is clean, accurate, and relevant before
training the AI model. Poor-quality data can lead to incorrect predictions and
unreliable AI models.
13. What is the role of feature engineering in AI?
Feature engineering transforms raw data into meaningful input variables that
improve model accuracy. Example: In fraud detection, creating a feature like "average
transaction amount per day" can help identify suspicious activity.
14. What is the difference between model training and model deployment?
 Model Training – The phase where the AI model learns from historical data.
 Model Deployment – The phase where the trained model is used in real-
world applications to make predictions.
15. Define overfitting in machine learning.
Overfitting occurs when an AI model memorizes the training data instead of
learning general patterns, resulting in poor performance on new, unseen data.
16. What is underfitting, and how does it affect AI models?
Underfitting happens when an AI model fails to capture patterns in data due to
being too simple, leading to low accuracy on both training and test data.
17. What is the purpose of model evaluation in AI?
Model evaluation helps measure how well the AI model performs using metrics like
RMSE, MSE, accuracy, and precision.
18. What is the difference between RMSE and MSE?
 RMSE (Root Mean Squared Error) – Measures error with more weight on
large errors.
 MSE (Mean Squared Error) – Measures error by averaging squared
differences but does not take the square root.
19. Why is cross-validation better than a simple train-test split?
Cross-validation ensures that the model is tested on multiple subsets of data,
reducing bias and improving performance evaluation.
20. What is K-Fold Cross-Validation, and how does it work?
K-Fold Cross-Validation divides the dataset into K equal parts. The model is trained
on K-1 parts and tested on the remaining part. This process is repeated K times, and
the results are averaged for better accuracy.

AI Evaluation Metrics
21. What is RMSE, and why is it used in model evaluation?
RMSE (Root Mean Squared Error) is a measure of how accurately a regression
model predicts values. It is useful because it penalizes large errors more heavily.
22. How is MSE different from MAE?
 MSE (Mean Squared Error) squares the differences, making it more sensitive
to large errors.
 MAE (Mean Absolute Error) takes the absolute differences, giving equal
weight to all errors.
23. What does Mean Absolute Error (MAE) measure?
MAE measures the average absolute difference between actual and predicted values,
making it easier to interpret.
24. Why do we use performance metrics in AI models?
Performance metrics help us assess how well an AI model is performing and
identify areas for improvement.
Descriptive Questions
1. Explain the AI project cycle in detail with an example.
o The AI project cycle consists of six steps:
 Problem Definition
 Data Gathering
 Feature Definition
 Model Construction
 Evaluation
 Deployment.
o Example: In a movie ticket price predictor project, problem definition
involves understanding factors influencing ticket prices, data gathering
includes collecting historical ticket sales, feature definition selects key
factors like location and show timing, model construction applies
 machine learning, evaluation assesses accuracy, and deployment
integrates the model into a real-world application.
2. Describe the five types of predictive analysis questions AI can answer.
o AI predictive analysis helps answer:
 Which category? (Classification) – e.g., Identifying spam
emails.
 How much or how many? (Regression) – e.g., Predicting house
prices.
 Which group? (Clustering) – e.g., Customer segmentation in
marketing.
 Is this unusual? (Anomaly Detection) – e.g., Fraud detection in
banking.
 Which option should be taken? (Recommendation) – e.g.,
Personalized movie suggestions.
3. How does Design Thinking help in AI problem-solving?
o Design Thinking is a solution-based approach that helps in solving
complex AI problems by breaking them into five stages:
1. Empathize: Understanding user needs.
2. Define: Clearly outlining the problem.
3. Ideate: Brainstorming potential AI solutions.
4. Prototype: Developing a small-scale version for testing.
5. Test: Evaluating and refining the model for real-world
application.
4. Discuss the role of data storytelling in AI and give an example.
o Data storytelling transforms raw data into meaningful narratives using
visuals and interpretation. It enhances decision-making by making
insights accessible and engaging.
o Example: A school uses data storytelling to analyze student engagement
in science subjects. Initially, 40% of students found science boring.
After implementing interactive teaching methods, a new survey revealed
a 38% increase in interest, demonstrating the impact of changes through
visual storytelling.
5. Compare Train-Test Split and Cross-Validation in AI model evaluation.
o Train-Test Split:
 Splits dataset into two parts (e.g., 80% training, 20% testing).
 Faster but may lead to overfitting if data is small.
o Cross-Validation:
 Divides data into multiple subsets and runs multiple
training/testing cycles.
 Provides more reliable performance estimates but is
computationally intensive.

Long Answer Type Questions

1. Explain the AI Model Life Cycle and its importance in AI projects.

The AI Model Life Cycle is a structured process followed for developing AI-based
solutions. It consists of three main stages:
1. Project Scoping:
o This involves defining the problem statement, identifying objectives,
and setting performance metrics.
o Example: A healthcare AI project aiming to predict heart disease must
define the input variables (e.g., patient age, lifestyle, medical history).
2. Model Building:
o Data is collected, cleaned, and prepared.
o Feature engineering is performed to extract meaningful insights from
raw data.
o The AI model is trained using machine learning algorithms.
o Model evaluation is done using metrics like RMSE, MSE, and accuracy
scores.
o Example: A fraud detection system in banking trains a model to
recognize unusual transaction patterns.
3. Deployment & Monitoring:
o The model is integrated into real-world applications.
o It is continuously monitored and updated to maintain accuracy.
o Example: A chatbot used in customer support is refined over time based
on user interactions.
Importance of AI Model Life Cycle:
 Ensures structured development of AI solutions.
 Helps in maintaining accuracy and efficiency of AI models.
 Supports continuous improvement by refining AI models based on real-world
data.

2. What is RMSE, and how is it used to measure AI model accuracy? Provide

an example.
Root Mean Squared Error (RMSE) is a statistical measure that evaluates how well a
machine learning model predicts outcomes compared to actual values. It calculates
the average magnitude of errors, where larger errors are given more weight.
Formula:-

Example:
Consider a model predicting house prices:
Actual Price (in $1000) Predicted Price (in $1000) Error

500 520 20

600 590 10

700 720 20

800 790 10
The RMSE value for this dataset would be calculated, and if it is low (e.g., 15), it
indicates a highly accurate model. A higher RMSE (e.g., 100) would suggest the
model needs improvement.
Why is RMSE useful?
 It penalizes large errors more heavily.
 It helps compare different machine learning models.
 Lower RMSE indicates better performance.

3. What is Cross-Validation? How does it improve AI model performance?

Cross-Validation (CV) is a technique used in machine learning to improve model
reliability by evaluating it on different subsets of data. Instead of a simple train-test
split, CV divides data into multiple sections to ensure a fair assessment of model
performance.
Types of Cross-Validation:
1. K-Fold Cross-Validation:
o Data is split into K parts (e.g., K=5 means data is divided into 5 sets).
o The model is trained on K-1 sets and tested on the remaining one.
o The process is repeated K times, and the final score is the average of all
iterations.
2. Leave-One-Out Cross-Validation (LOOCV):
o Each instance in the dataset is used as a test case once, while all others
are used for training.
o Best for small datasets but computationally expensive.
Example:
Imagine a dataset with 1000 records used in 5-Fold Cross-Validation. The model
will be trained and tested five times, each time using 800 records for training and
200 for testing.
Advantages of Cross-Validation:
 Provides a more reliable evaluation compared to a single train-test split.
 Reduces the chances of overfitting (where the model memorizes training data
instead of learning general patterns).
 Ensures the model performs well across different datasets.

4. What is Design Thinking, and how is it applied in AI development?

Design Thinking is a human-centered problem-solving approach that helps AI
developers create solutions that align with user needs. It consists of five stages:
1. Empathize:
 o Understand the user’s challenges and motivations.
o Example: Interview doctors before designing an AI system for medical
diagnosis.
2. Define:
o Clearly state the problem that needs to be solved.
o Example: "How can AI improve early detection of lung cancer?"
3. Ideate:
o Brainstorm multiple AI solutions and select the most feasible one.
o Example: Developing an AI-based image recognition tool for detecting
lung cancer in X-ray images.
4. Prototype:
o Build a basic version of the AI system and test it.
o Example: Create a machine learning model that classifies X-ray
images as "normal" or "abnormal."
5. Test:
o Validate the model using real-world data and refine it based on
feedback.
o Example: Train the AI model with thousands of X-rays and adjust its
algorithm for better accuracy.
Why is Design Thinking important in AI?
 Ensures AI solutions are user-friendly and problem-oriented.
 Encourages iterative improvement, making AI models more effective over
time.
 Helps in addressing ethical concerns by focusing on user impact.

5. How does Data Storytelling improve AI decision-making? Give an example.

Data Storytelling is the art of presenting data insights in a narrative format using
visuals, context, and interpretation to make complex information easy to
understand.
Elements of Data Storytelling:
1. Understanding the Audience: Tailor the story to their knowledge level.
2. Choosing the Right Data & Visualizations: Use charts, graphs, and
infographics.
3. Drawing Attention to Key Information: Highlight trends and patterns.
4. Developing a Narrative: Explain why data changes over time.
5. Engaging the Audience: Make the insights compelling and actionable.
Example:
A teacher collects feedback from students about their interest in Science. The pre-
survey results show:
 40% of students feel bored in class.
 Only 19% feel excited about Science.
After changing the teaching approach, a post-survey reveals:
 38% of students now feel excited about Science.
 Only 12% feel bored.
By using charts and a compelling story, the teacher can present the data
effectively and justify changes in teaching methods.
Why is Data Storytelling important?
 Helps decision-makers understand why patterns emerge in data.
 Makes AI insights more relatable and actionable.
 Reduces data misinterpretation by providing clear context.