DATA SCIENCE (AI)

OBJECTIVE QUESTIONS

1. What is the primary role of Data Science in Artificial Intelligence (AI)?

(a) Designing user interfaces
(b) Unifying statistics, data analysis, and machine learning
(c) Building computer hardware
(d) Enhancing battery life in AI-powered devices
Answer: (b) Unifying statistics, data analysis, and machine learning
Explanation: Data Science combines various techniques from statistics, data analysis, and machine
learning to analyze and interpret data, making AI systems intelligent.

2. Which of the following is NOT a domain of AI based on data type?

(a) Computer Vision (CV)
(b) Natural Language Processing (NLP)
(c) Cybersecurity
(d) Data Science
Answer: (c) Cybersecurity
Explanation: AI is classified into three main domains: Data Science (numeric and alpha-numeric
data), Computer Vision (image and visual data), and Natural Language Processing (text and speech
data). Cybersecurity is an application area but not a core AI domain.

3. What was the earliest application of Data Science in the finance industry?
(a) Fraud and Risk Detection
(b) Customer Service Automation
(c) Cryptocurrency Trading
(d) ATM Cash Dispensing
Answer: (a) Fraud and Risk Detection
Explanation: Data Science was initially used in finance for fraud and risk detection, helping banks
analyze customer data to reduce losses from bad debts and defaults.

4. How do search engines like Google use Data Science?

(a) By manually ranking webpages
(b) By using data science algorithms to retrieve relevant search results
(c) By collecting physical copies of web pages
(d) By randomly displaying search results
Answer: (b) By using data science algorithms to retrieve relevant search results
Explanation: Search engines process vast amounts of data using advanced algorithms to provide
accurate and relevant search results within seconds.

5. Which of the following is NOT a commonly used file format for storing structured data?
(a) CSV
(b) SQL
(c) PNG
(d) Spreadsheet
Answer: (c) PNG
Explanation: PNG is an image file format, whereas CSV, SQL, and Spreadsheet formats are
commonly used for storing and managing structured data.

6. What is the main advantage of digital advertisements using Data Science?

(a) They are always free to display
(b) They are randomly placed on websites

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 1 -

(c) They have a higher Call-Through Rate (CTR) due to targeted recommendations
(d) They require manual user tracking
Answer: (c) They have a higher Call-Through Rate (CTR) due to targeted recommendations
Explanation: Digital ads utilize data science algorithms to analyze user behavior and display
personalized ads, increasing their effectiveness.

7. In a Data Science project, why is the AI Project Cycle important?

(a) It eliminates the need for human intervention
(b) It ensures that AI models are built systematically, from problem scoping to evaluation
(c) It only focuses on collecting data without analysis
(d) It prevents the AI model from making predictions
Answer: (b) It ensures that AI models are built systematically, from problem scoping to evaluation
Explanation: The AI Project Cycle provides a structured approach to solving problems using AI,
ensuring each step is carefully planned and executed.

8. Which Python library is primarily used for numerical computations in Data Science?
(a) Pandas
(b) NumPy
(c) Matplotlib
(d) Seaborn
Answer: (b) NumPy
Explanation: NumPy is a core library for numerical computing in Python, providing support for
large multidimensional arrays and mathematical functions.

9. In the K-Nearest Neighbors (KNN) algorithm, what does the variable ‘K’ represent?
(a) The number of nearest neighbors considered for classification
(b) The number of layers in a neural network
(c) The total number of data points in the dataset
(d) The maximum distance allowed between data points
Answer: (a) The number of nearest neighbors considered for classification
Explanation: In KNN, ‘K’ determines how many closest data points are used to classify a new data
point, affecting the accuracy of predictions.

10. What is the main purpose of using Matplotlib in Data Science?

(a) To create and visualize data using various types of graphs
(b) To clean datasets by removing missing values
(c) To store large datasets efficiently
(d) To perform machine learning model training
Answer: (a) To create and visualize data using various types of graphs
Explanation: Matplotlib is a Python library used for data visualization, helping analysts interpret
trends and patterns in data more effectively.

11. Which of the following is NOT an application of Data Science?

(a) Fraud and Risk Detection
(b) Targeted Advertising
(c) Genetic Research
(d) Manual Data Entry
Answer: (d) Manual Data Entry
Explanation: Data Science automates data processing and analysis. Manual data entry is a human-
driven process and does not utilize data science techniques.

12. What is the main objective of a regression model in Data Science?

(a) To classify data into categories
(b) To generate random numbers

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 2 -

(c) To predict continuous numerical values based on historical data
(d) To identify images in a dataset
Answer: (c) To predict continuous numerical values based on historical data
Explanation: Regression models are used in Data Science to predict numerical values, such as sales
forecasts or temperature predictions, based on past data trends.

13. Which of the following data types is mainly used in Natural Language Processing (NLP)?
(a) Image data
(b) Text and speech data
(c) Numeric data
(d) Sensor data
Answer: (b) Text and speech data
Explanation: NLP deals with processing human language in text or speech format to enable AI to
understand, interpret, and generate responses.

14. What does CSV stand for in Data Science?

(a) Computer Standard Value
(b) Comma Separated Values
(c) Coded System Variable
(d) Calculated Statistical Variance
Answer: (b) Comma Separated Values
Explanation: CSV is a widely used file format for storing tabular data where each value is separated
by a comma, making it easy to handle structured datasets.

15. Why is data visualization important in Data Science?

(a) It helps machines learn faster
(b) It makes large amounts of data easier to interpret for humans
(c) It replaces statistical analysis
(d) It prevents data errors
Answer: (b) It makes large amounts of data easier to interpret for humans
Explanation: Data visualization tools like Matplotlib and Seaborn help present data graphically,
making patterns, trends, and insights easier to understand.

16. In Data Science, what is the purpose of cleaning data?

(a) To remove unnecessary data for faster storage
(b) To make the dataset error-free and more accurate
(c) To convert data into images
(d) To increase the file size
Answer: (b) To make the dataset error-free and more accurate
Explanation: Data cleaning ensures that missing values, duplicates, and incorrect data entries are
handled properly, improving the quality of analysis and model predictions.

17. How does targeted advertising use Data Science?

(a) By randomly showing advertisements to users
(b) By using user data to display relevant ads based on interests and behavior
(c) By advertising only in newspapers
(d) By increasing the price of digital ads
Answer: (b) By using user data to display relevant ads based on interests and behavior
Explanation: Data Science helps in analyzing user behavior, preferences, and past interactions to
deliver highly relevant advertisements, improving engagement and conversions.

18. Which algorithm is commonly used for classification tasks in Data Science?
(a) Linear Regression
(b) K-Nearest Neighbors (KNN)

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 3 -

(c) K-Means Clustering
(d) Principal Component Analysis (PCA)
Answer: (b) K-Nearest Neighbors (KNN)
Explanation: KNN is a supervised machine learning algorithm used for classification tasks. It
assigns a class to a new data point based on the majority class of its nearest neighbors.

19. What is the significance of outliers in a dataset?

(a) They are always errors and should be removed
(b) They can provide valuable insights but may also indicate data errors
(c) They make the dataset invalid
(d) They increase model accuracy
Answer: (b) They can provide valuable insights but may also indicate data errors
Explanation: Outliers are extreme values in a dataset. While they may indicate errors, they can also
reveal hidden trends or rare events that need further analysis.

20. What is the main function of the Pandas library in Python?

(a) Creating 3D graphics
(b) Performing complex mathematical calculations
(c) Handling and analyzing structured data efficiently
(d) Controlling hardware devices
Answer: (c) Handling and analyzing structured data efficiently
Explanation: Pandas provides powerful data structures such as DataFrames and Series to manipulate
and analyze tabular data effectively.

21. What is the main reason airlines use Data Science?

(a) To increase in-flight entertainment options
(b) To predict flight delays and optimize routes
(c) To manufacture airplanes
(d) To monitor passenger emotions
Answer: (b) To predict flight delays and optimize routes
Explanation: Airlines use Data Science to analyze flight delays, decide aircraft types, plan routes,
and improve customer loyalty programs, helping optimize operations.

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

(a) Model Training
(b) Data Collection
(c) Problem Scoping
(d) Model Evaluation
Answer: (c) Problem Scoping
Explanation: Problem Scoping involves defining the issue, identifying stakeholders, and
understanding the problem’s context before collecting data and training an AI model.

23. Which of the following Python libraries is specifically used for data visualization?
(a) Pandas
(b) NumPy
(c) Matplotlib
(d) Scikit-learn
Answer: (c) Matplotlib
Explanation: Matplotlib is a Python library used for data visualization, helping to create graphs,
charts, and plots for better understanding of data patterns.

24. How does Data Science contribute to personalized medicine in healthcare?

(a) By recommending random treatments
(b) By analyzing genetic data to predict drug responses and diseases

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 4 -

(c) By replacing doctors with AI robots
(d) By providing general medical advice to all patients
Answer: (b) By analyzing genetic data to predict drug responses and diseases
Explanation: Data Science is used in genetics and genomics to understand the relationship between
DNA and diseases, leading to personalized treatment plans.

25. What does the term "Big Data" refer to in Data Science?
(a) Data that is too large and complex to be processed using traditional methods
(b) Data stored on personal computers
(c) Data that is always stored in physical files
(d) Small datasets that fit within a spreadsheet
Answer: (a) Data that is too large and complex to be processed using traditional methods
Explanation: Big Data involves large volumes of structured and unstructured data that require
advanced tools and techniques to process and analyze effectively.

26. Which method is commonly used to handle missing values in a dataset?

(a) Ignoring the missing values
(b) Replacing them with mean, median, or mode
(c) Deleting the entire dataset
(d) Changing them to random values
Answer: (b) Replacing them with mean, median, or mode
Explanation: Missing values are often replaced with statistical measures like mean, median, or mode
to maintain the integrity of the dataset without losing valuable information.

27. In Data Science, what does "training a model" mean?

(a) Teaching the model how to play a game
(b) Feeding data to the model so it can learn patterns and make predictions
(c) Printing a dataset for future reference
(d) Making the AI memorize all possible outputs
Answer: (b) Feeding data to the model so it can learn patterns and make predictions
Explanation: Training a model involves providing it with a dataset so it can identify patterns and
make accurate predictions on new, unseen data.

28. Why do companies use recommendation systems?

(a) To randomly suggest products to customers
(b) To offer personalized product or content suggestions based on user preferences
(c) To prevent users from purchasing items
(d) To reduce the number of options available to customers
Answer: (b) To offer personalized product or content suggestions based on user preferences
Explanation: Companies like Amazon, Netflix, and Spotify use recommendation systems to analyze
user behavior and suggest relevant products or content.

29. What is the key feature of supervised learning in Data Science?

(a) It does not require labeled data
(b) The model learns from labeled training data
(c) It only works with images
(d) It randomly classifies data without any learning process
Answer: (b) The model learns from labeled training data
Explanation: In supervised learning, the model is trained using labeled data, meaning it learns from
input-output pairs to make accurate predictions.

30. What is an advantage of using Pandas in Data Science?

(a) It speeds up image recognition
(b) It provides easy-to-use data structures for handling and analyzing structured data

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 5 -

(c) It replaces the need for Python programming
(d) It is used for hardware programming
Answer: (b) It provides easy-to-use data structures for handling and analyzing structured data
Explanation: Pandas is a powerful Python library that helps in handling, analyzing, and manipulating
structured data efficiently using DataFrames and Series.

31. Which of the following is NOT one of the three broad AI domains discussed in the
document?
(a) Data Sciences
(b) Computer Vision (CV)
(c) Natural Language Processing (NLP)
(d) Robotics
Answer: (d) Robotics
Explanation: The document categorizes AI into three main domains based on the type of data used:
Data Sciences (numeric/alpha-numeric), Computer Vision (image/visual), and Natural Language
Processing (text/speech). Robotics is not listed as one of these core domains.

32. In the context of AI domains, what type of data does the Computer Vision domain primarily
work with?
(a) Textual data
(b) Numerical data
(c) Image and visual data
(d) Speech data
Answer: (c) Image and visual data
Explanation: Computer Vision (CV) focuses on processing and analyzing image and visual data,
distinguishing it from other domains that work with text, numbers, or speech.

33. Which file format is described as a simple, text-based format where each line is a record and
fields are separated by commas?
(a) SQL
(b) Spreadsheet
(c) CSV
(d) XML
Answer: (c) CSV
Explanation: CSV stands for Comma Separated Values. It is a plain text format used for storing
tabular data, where each record is on a new line and individual values are separated by commas.

34. What is a key advantage of using NumPy arrays over Python lists?
(a) They can hold heterogeneous data types
(b) They allow direct element-wise arithmetic operations
(c) They use more memory than lists
(d) They can be directly initialized using Python syntax
Answer: (b) They allow direct element-wise arithmetic operations
Explanation: NumPy arrays are optimized for numerical computations; they support direct
element-wise operations (like division of all elements by a constant) and are more memory-efficient
compared to Python lists, which are designed for general data management.

35. Which statistical measure is most affected by extreme values (outliers) in a dataset?
(a) Median
(b) Mode
(c) Mean
(d) Interquartile Range
Answer: (c) Mean

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 6 -

Explanation: The mean (average) is sensitive to extreme values because it takes into account every
value in the dataset. Outliers can skew the mean significantly, unlike the median or mode.

36. In a box plot, what does the interquartile range (IQR) represent?
(a) The difference between the maximum and minimum values
(b) The range between the 25th and 75th percentiles
(c) The average value of the dataset
(d) The sum of all quartile values
Answer: (b) The range between the 25th and 75th percentiles
Explanation: The interquartile range (IQR) measures the middle 50% of the data by calculating the
difference between the 75th percentile (upper quartile) and the 25th percentile (lower quartile),
providing insight into the data’s spread while mitigating the effect of outliers.

37. Which method is a classic example of offline data collection as mentioned in the document?
(a) Sensors
(b) Surveys
(c) Open-sourced Government Portals
(d) Reliable websites
Answer: (b) Surveys
Explanation: Offline data collection refers to methods like surveys, interviews, and observations.
Surveys are conducted in-person or on paper, in contrast to online methods that use digital portals or
websites.

38. What is the primary role of the Pandas library in Python for Data Science?
(a) Creating visual plots and graphs
(b) Performing advanced numerical computations on multi-dimensional arrays
(c) Handling and analyzing structured (tabular) data
(d) Web scraping
Answer: (c) Handling and analyzing structured (tabular) data
Explanation: Pandas provides powerful data structures such as DataFrames and Series, which are
designed to manipulate, analyze, and manage structured data—making it essential for data analysis
tasks.

39. Which type of plot is most suitable for visualizing the distribution of data along with its
quartiles and outliers?
(a) Scatter plot
(b) Bar chart
(c) Box plot
(d) Line chart
Answer: (c) Box plot
Explanation: A box plot (or box-and-whiskers plot) is designed to display the distribution of data by
showing the median, quartiles, and potential outliers, offering a clear view of data variability.

40. In the K-Nearest Neighbors (KNN) algorithm, the prediction for a new data point is based
on which of the following?
(a) The global distribution of the dataset
(b) A random selection of data points
(c) The labels of the closest data points in the feature space
(d) Predefined thresholds set by the user
Answer: (c) The labels of the closest data points in the feature space
Explanation: KNN relies on the assumption that similar data points exist close to each other in the
feature space. It classifies a new point by considering the labels of its K nearest neighbors and
choosing the majority vote (or average, for regression).

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 7 -

41. What is the primary goal of Data Science in AI?
(a) To replace human intelligence completely
(b) To unify statistics, data analysis, and machine learning to analyze phenomena
(c) To store large amounts of data without processing it
(d) To create random datasets
Answer: (b) To unify statistics, data analysis, and machine learning to analyze phenomena
Explanation: Data Science combines various fields, such as statistics, data analysis, and machine
learning, to extract meaningful insights and make AI systems more efficient.

42. Which industry was among the first to implement Data Science for practical applications?
(a) Healthcare
(b) Finance
(c) Entertainment
(d) Agriculture
Answer: (b) Finance
Explanation: The finance industry was one of the earliest adopters of Data Science, using it for fraud
detection, risk assessment, and customer profiling.

43. In a dataset, which statistical measure helps in identifying the most frequently occurring
value?
(a) Mean
(b) Median
(c) Mode
(d) Standard Deviation
Answer: (c) Mode
Explanation: The mode is the value that appears most frequently in a dataset, making it useful for
identifying common trends or repeated patterns.

44. Which of the following is an example of structured data?

(a) A collection of social media posts
(b) A set of customer reviews in text format
(c) A spreadsheet containing employee salaries
(d) A set of audio recordings from customer calls
Answer: (c) A spreadsheet containing employee salaries
Explanation: Structured data is organized in a predefined format, such as tables with rows and
columns, making it easier to analyze using database systems.

45. What does a histogram primarily show?

(a) The relationship between two categorical variables
(b) The frequency distribution of a continuous variable
(c) The correlation between two variables
(d) The cumulative sum of values in a dataset
Answer: (b) The frequency distribution of a continuous variable
Explanation: A histogram represents the distribution of a continuous variable by dividing it into
intervals (bins) and displaying the frequency of observations in each interval.

46. Which of the following sources can be used for online data collection in Data Science?
(a) Interviews
(b) Open-sourced government portals
(c) Paper-based surveys
(d) Personal conversations
Answer: (b) Open-sourced government portals
Explanation: Online data collection sources include publicly available government portals, reliable
websites, and statistical databases that provide structured and authenticated datasets.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 8 -

47. What is the primary purpose of regression analysis in Data Science?
(a) To categorize data into different groups
(b) To determine relationships between variables and predict numerical outcomes
(c) To clean large datasets for better visualization
(d) To detect anomalies in categorical data
Answer: (b) To determine relationships between variables and predict numerical outcomes
Explanation: Regression analysis is used to understand the relationship between variables and make
numerical predictions based on historical data trends.

48. Why is data cleaning an essential step in the Data Science process?
(a) To reduce the size of the dataset
(b) To remove irrelevant data that may affect the model’s accuracy
(c) To improve internet connectivity while accessing databases
(d) To convert all numerical values into text format
Answer: (b) To remove irrelevant data that may affect the model’s accuracy
Explanation: Data cleaning helps in removing errors, inconsistencies, and irrelevant information,
ensuring that the dataset used for analysis or machine learning models is accurate and reliable.

49. Which visualization technique is most suitable for showing trends over time?
(a) Pie chart
(b) Scatter plot
(c) Line graph
(d) Box plot
Answer: (c) Line graph
Explanation: Line graphs are ideal for visualizing data over time, helping to identify trends, patterns,
and fluctuations in a dataset.

50. In the AI project cycle, what comes immediately after defining the problem?
(a) Data Collection
(b) Model Deployment
(c) Model Training
(d) Data Visualization
Answer: (a) Data Collection
Explanation: Once a problem is identified, the next step in the AI project cycle is collecting relevant
data to analyze and build a model that can provide solutions.

51. What is the primary function of a classification model in Data Science?

(a) To predict numerical values
(b) To group data into predefined categories
(c) To generate random data points
(d) To store large amounts of data
Answer: (b) To group data into predefined categories
Explanation: Classification models are used in machine learning to assign data points to specific
categories, such as spam detection in emails or disease diagnosis.

52. What does the K in K-Nearest Neighbors (KNN) represent?

(a) The number of nearest neighbors used for classification
(b) The total number of data points in a dataset
(c) The accuracy of a machine learning model
(d) The size of the dataset
Answer: (a) The number of nearest neighbors used for classification
Explanation: In the KNN algorithm, K represents the number of closest data points considered when
classifying a new data point.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 9 -

53. What is an outlier in a dataset?
(a) A missing value
(b) A data point significantly different from the rest
(c) A duplicate record in a dataset
(d) A randomly generated number
Answer: (b) A data point significantly different from the rest
Explanation: Outliers are values that deviate significantly from other observations, which can affect
statistical analysis and machine learning models.

54. Which Python library is primarily used for handling structured data in tabular form?
(a) Matplotlib
(b) NumPy
(c) Pandas
(d) TensorFlow
Answer: (c) Pandas
Explanation: Pandas is a Python library designed for working with structured data using data
structures like DataFrames and Series.

55. Why is data visualization important in Data Science?

(a) It speeds up machine learning model training
(b) It helps humans understand complex datasets by representing data graphically
(c) It reduces the file size of a dataset
(d) It replaces the need for data preprocessing
Answer: (b) It helps humans understand complex datasets by representing data graphically
Explanation: Data visualization tools like Matplotlib and Seaborn help in presenting data trends,
patterns, and insights in an easily understandable format.

56. Which of the following is NOT an application of Data Science?

(a) Fraud detection
(b) Weather forecasting
(c) Mobile phone charging
(d) Customer recommendation systems
Answer: (c) Mobile phone charging
Explanation: While Data Science is used in fraud detection, weather forecasting, and
recommendation systems, it is not directly related to charging mobile phones.

57. What type of dataset does a regression model use?

(a) Only categorical data
(b) Only text data
(c) Continuous numerical data
(d) Image data
Answer: (c) Continuous numerical data
Explanation: Regression models predict continuous numerical values, such as sales forecasts or
temperature predictions, based on historical data.

58. Which type of Data Science model would be used for predicting whether an email is spam or
not?
(a) Regression model
(b) Classification model
(c) Clustering model
(d) Reinforcement learning model
Answer: (b) Classification model

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 10 -

Explanation: Classification models are used to assign labels to data, such as categorizing emails as
spam or non-spam.

59. What is the purpose of feature selection in Data Science?

(a) To increase the number of features in a dataset
(b) To select the most relevant features for improving model performance
(c) To remove all numerical data from a dataset
(d) To convert text data into numerical form
Answer: (b) To select the most relevant features for improving model performance
Explanation: Feature selection helps improve the accuracy and efficiency of machine learning
models by reducing unnecessary or redundant data.

60. What is the primary goal of anomaly detection in Data Science?

(a) To identify unusual patterns in data
(b) To delete duplicate records in a dataset
(c) To increase dataset size
(d) To visualize numerical data
Answer: (a) To identify unusual patterns in data
Explanation: Anomaly detection is used to find data points that do not conform to expected patterns,
such as fraud detection or network intrusion detection.

QUESTIONS AND ANSWERS (2 marks)

1. What are the three main domains of AI based on data types?

Answer:
The three main domains of AI based on data types are:
1. Data Science – Works with numeric and alpha-numeric data.
2. Computer Vision (CV) – Deals with image and visual data.
3. Natural Language Processing (NLP) – Handles textual and speech-based data.
Explanation:
AI is classified based on the type of data it processes. Data Science deals with structured data, CV
focuses on visual inputs, and NLP enables AI to understand human language.

2. Why is Data Science important in AI?

Answer:
Data Science helps AI systems analyze, interpret, and make decisions based on data. It enables AI
models to learn patterns, predict outcomes, and automate decision-making.
Explanation:
AI is heavily reliant on data. Without Data Science, AI models would lack the ability to extract
insights, making them ineffective in applications like fraud detection, recommendation systems, and
automation.

3. What are some key applications of Data Science?

Answer:
Some key applications of Data Science include:
 Fraud Detection – Identifying suspicious activities in banking and finance.
 Search Engine Optimization – Google and Bing use data science for fast and relevant
searches.
 Recommendation Systems – Netflix, Amazon, and YouTube suggest content based on user
preferences.
 Healthcare Analytics – Analyzing patient records to predict diseases.
Explanation:
Data Science is widely used in different sectors to improve efficiency, detect patterns, and provide
better services through AI-powered solutions.
Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 11 -
4. How do search engines like Google use Data Science?
Answer:
Search engines use Data Science algorithms to index web pages, rank results, and provide the most
relevant information to users within milliseconds.
Explanation:
Search engines process huge amounts of data daily and use machine learning algorithms to rank
and refine search results based on factors like keywords, user history, and webpage relevance.

5. What is the AI Project Cycle and why is it important?

Answer:
The AI Project Cycle consists of the following steps:
1. Problem Scoping – Identifying the problem to be solved.
2. Data Collection – Gathering relevant data.
3. Data Processing – Cleaning and structuring data.
4. Model Training – Training AI models using collected data.
5. Model Evaluation – Checking model accuracy and performance.
Explanation:
The AI Project Cycle ensures a structured approach to building AI models, helping teams
systematically develop and deploy solutions with high accuracy.

6. What are the different types of data formats used in Data Science?
Answer:
The common data formats used in Data Science are:
1. CSV (Comma Separated Values) – Stores tabular data in a simple text format.
2. Spreadsheet (Excel files) – Used for organizing and analyzing structured data.
3. SQL Databases – Stores and manages large datasets efficiently.
Explanation:
Different formats are used depending on the requirement. CSV is easy to use, spreadsheets allow for
manual analysis, and SQL databases are suitable for handling large-scale structured data.

7. What is the difference between structured and unstructured data?

Answer:
 Structured Data – Organized in rows and columns, such as spreadsheets and databases.
 Unstructured Data – Lacks a predefined structure, such as images, videos, and social media
posts.
Explanation:
Structured data is easy to analyze using traditional database systems, whereas unstructured data
requires advanced AI techniques like Computer Vision and NLP for processing.

8. How does Data Science help in fraud detection?

Answer:
Data Science helps in fraud detection by analyzing customer transactions, identifying unusual
patterns, and flagging suspicious activities for further review.
Explanation:
Banks and financial institutions use machine learning algorithms to detect anomalies in spending
patterns. AI models learn from past fraud cases and apply this knowledge to real-time transactions to
prevent fraud.

9. What are the advantages of using Python for Data Science?

Answer:
Python is preferred for Data Science because:
1. Rich Libraries – Offers powerful libraries like NumPy, Pandas, and Matplotlib.
2. Easy Syntax – Simple and readable syntax makes it user-friendly.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 12 -

 3. Community Support – Large open-source community provides extensive resources.
Explanation:
Python simplifies data manipulation, visualization, and machine learning, making it an ideal choice
for Data Science applications.

10. What is meant by data preprocessing in Data Science?

Answer:
Data preprocessing refers to the steps taken to clean and prepare data before analysis. It includes:
 Removing duplicates
 Handling missing values
 Normalizing and transforming data
Explanation:
Preprocessing ensures that datasets are accurate and well-structured, leading to better model
performance. Without it, AI models may produce incorrect or misleading results.

11. What are the two main types of data collection methods in Data Science?
Answer:
1. Offline Data Collection – Includes surveys, interviews, and observations.
2. Online Data Collection – Includes web scraping, open government portals, and databases
like Kaggle.
Explanation:
Data Science relies on high-quality data, which can be collected either manually (offline) or from
digital sources (online). The choice of method depends on the availability and reliability of data.

12. What is the role of NumPy in Data Science?

Answer:
NumPy (Numerical Python) is used for:
 Performing mathematical operations on large datasets.
 Handling multi-dimensional arrays efficiently.
Explanation:
NumPy is essential in Data Science because it provides fast and efficient array operations, making
data analysis and machine learning tasks computationally efficient.

13. How does Data Science contribute to targeted advertising?

Answer:
Data Science analyzes user behavior, preferences, and past interactions to display personalized ads
that match user interests.
Explanation:
Platforms like Google, Facebook, and Amazon use machine learning algorithms to predict which
ads will be most relevant to each user, improving engagement and sales.

14. What are the different types of Machine Learning models used in Data Science?
Answer:
1. Supervised Learning – Trained on labeled data (e.g., classification and regression).
2. Unsupervised Learning – Identifies patterns in unlabeled data (e.g., clustering).
3. Reinforcement Learning – Learns through rewards and penalties (e.g., robotics, game
playing).
Explanation:
Different machine learning techniques are applied based on the nature of the data and the problem to
be solved, ensuring accurate and efficient predictions.

15. What is the difference between Mean and Median in statistics?

Answer:
 Mean – The average of all numbers in a dataset.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 13 -

  Median – The middle value when the numbers are arranged in order.
Explanation:
Mean is sensitive to outliers, which can distort the result, whereas the median gives a better central
value when dealing with skewed data.

16. Why is data cleaning important in Data Science?

Answer:
Data cleaning removes incorrect, incomplete, or irrelevant data, improving model accuracy and
performance.
Explanation:
Poor-quality data can lead to misleading results. Cleaning ensures that data is consistent, complete,
and free of errors, which is essential for effective analysis and AI model training.

17. How does a recommendation system work in Data Science?

Answer:
A recommendation system uses past user behavior to suggest relevant products, movies, or content.
Explanation:
Companies like Amazon, Netflix, and Spotify use collaborative filtering and content-based
filtering to provide personalized recommendations, improving user engagement and experience.

18. What is feature engineering in Data Science?

Answer:
Feature engineering is the process of selecting and transforming raw data into meaningful features
that improve machine learning model performance.
Explanation:
Good feature engineering helps models recognize patterns more efficiently, leading to higher
accuracy and better predictions in AI applications.

19. What are histograms used for in Data Science?

Answer:
Histograms are used to visualize the frequency distribution of a continuous dataset.
Explanation:
By grouping data into bins, histograms help identify trends, patterns, and outliers, making them
useful in exploratory data analysis.

20. How does Data Science help in airline route planning?

Answer:
Airlines use Data Science to:
 Predict flight delays.
 Optimize fuel efficiency and routes.
 Improve customer experience through loyalty programs.
Explanation:
By analyzing historical data, weather conditions, and demand patterns, airlines make data-driven
decisions to reduce costs and enhance efficiency.

21. What are the key differences between classification and regression models?
Answer:
 Classification Models predict categorical outputs (e.g., "spam" or "not spam").
 Regression Models predict continuous numerical values (e.g., temperature, sales).
Explanation:
Classification models are used for problems with distinct categories, while regression models deal
with numerical predictions based on input variables.

22. How does Data Science help in fraud detection?

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 14 -

Answer:
Data Science analyzes customer transaction patterns to identify suspicious activities and detect
potential fraud.
Explanation:
AI models learn from past fraudulent transactions and use anomaly detection algorithms to flag
unusual patterns in real time, preventing financial losses.

23. What is the role of Pandas in Data Science?

Answer:
Pandas is a Python library used for handling and analyzing structured data using DataFrames and
Series.
Explanation:
Pandas provides efficient tools for data manipulation, including filtering, merging, and statistical
analysis, making it essential for data preprocessing.

24. Why is feature selection important in Machine Learning?

Answer:
Feature selection removes irrelevant or redundant data, improving model performance and reducing
computation time.
Explanation:
Choosing only the most relevant features ensures that models learn efficiently, preventing overfitting
and enhancing accuracy.

25. What is the importance of exploratory data analysis (EDA) in Data Science?
Answer:
EDA helps in understanding patterns, detecting anomalies, and summarizing dataset
characteristics before model training.
Explanation:
EDA involves visualizing and statistically analyzing data to gain insights, ensuring that models are
trained on clean and meaningful data.

26. How does the K-Nearest Neighbors (KNN) algorithm work?

Answer:
KNN classifies a data point based on the majority class of its K nearest neighbors.
Explanation:
KNN is a supervised learning algorithm that assigns labels to new data points by calculating the
distance from existing labeled data, commonly using Euclidean distance.

27. What is the difference between correlation and causation in Data Science?
Answer:
 Correlation means two variables move together but may not be directly related.
 Causation means one variable directly affects the other.
Explanation:
A strong correlation does not imply causation. For example, an increase in ice cream sales correlates
with an increase in drowning cases, but eating ice cream does not cause drowning—summer is the
underlying factor.

28. What are outliers, and how do they impact a dataset?

Answer:
Outliers are extreme values that deviate significantly from the rest of the dataset, which can distort
statistical analysis and model performance.
Explanation:
Outliers can skew averages (mean), affect standard deviation, and lead to incorrect conclusions in
machine learning models. Identifying and handling them is crucial for accurate analysis.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 15 -

29. How does sentiment analysis work in Natural Language Processing (NLP)?
Answer:
Sentiment analysis detects emotions in text by classifying it as positive, negative, or neutral using
NLP techniques.
Explanation:
AI models use text mining and machine learning to analyze words, tone, and context in customer
reviews, social media posts, and surveys to gauge public opinion.

30. What is the significance of data visualization in Data Science?

Answer:
Data visualization transforms raw data into graphs, charts, and plots to help in understanding
patterns and trends.
Explanation:
Tools like Matplotlib and Seaborn provide insights by making large datasets easier to interpret,
assisting in decision-making and storytelling with data.

31. What is the difference between Supervised and Unsupervised Learning?

Answer:
 Supervised Learning: The model is trained on labeled data (e.g., classification, regression).
 Unsupervised Learning: The model finds patterns in unlabeled data (e.g., clustering,
anomaly detection).
Explanation:
Supervised learning requires input-output pairs for training, whereas unsupervised learning identifies
hidden structures without predefined labels.

32. What is the purpose of Normalization in Data Preprocessing?

Answer:
Normalization scales numerical data to a standard range (0 to 1 or -1 to 1), making features
comparable.
Explanation:
It improves machine learning model performance by preventing features with large numerical
values from dominating those with smaller values.

33. How does Data Science help in Healthcare?

Answer:
Data Science is used in disease prediction, medical image analysis, and personalized treatment
recommendations based on patient data.
Explanation:
By analyzing large datasets from genetics, medical records, and drug trials, AI helps doctors make
data-driven decisions and improve patient care.

34. What is an AI Model’s Training and Testing Process?

Answer:
1. Training Phase – The model learns from a dataset with labeled inputs and outputs.
2. Testing Phase – The trained model is evaluated on new data to measure accuracy.
Explanation:
A dataset is usually split (e.g., 80% training, 20% testing) to ensure that the model generalizes well
to unseen data.

35. What are the advantages of using Python for Data Science?
Answer:
Python is widely used because it has:
 Libraries like NumPy, Pandas, and Matplotlib for data analysis and visualization.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 16 -

  Easy syntax, making it beginner-friendly.
Explanation:
Python simplifies data manipulation, visualization, and machine learning model
implementation, making it ideal for Data Science.

36. What is the role of a Confusion Matrix in classification models?

Answer:
A confusion matrix evaluates model performance by showing True Positives, True Negatives, False
Positives, and False Negatives.
Explanation:
It helps in calculating important metrics like accuracy, precision, recall, and F1-score, which assess
a classification model’s effectiveness.

37. How does Data Science help in Weather Prediction?

Answer:
Weather prediction models use historical climate data and real-time sensor inputs to forecast
temperature, rainfall, and storms.
Explanation:
Using machine learning and statistical models, meteorologists analyze patterns in weather data to
improve forecasting accuracy.

38. What is Cross-Validation in Machine Learning?

Answer:
Cross-validation splits data into multiple training and testing sets to ensure model reliability and
reduce overfitting.
Explanation:
The K-Fold Cross-Validation method divides data into K subsets, training the model on K-1 parts
and testing on the remaining part, ensuring robustness.

39. What is Big Data, and why is it important?

Answer:
Big Data refers to large, complex datasets that require advanced processing techniques.
Explanation:
Big Data helps in business decision-making, real-time analytics, and AI model training, enabling
companies to gain insights from vast amounts of data.

40. How does Airline Route Optimization use Data Science?

Answer:
Airlines use machine learning models to optimize flight routes, reduce delays, and improve fuel
efficiency.
Explanation:
By analyzing past flight data, weather conditions, and passenger demand, Data Science helps
airlines operate more efficiently and minimize costs.

QUESTIONS AND ANSWERS - 3 marks

1. What are the main stages of the AI Project Cycle, and why is each stage important?
Answer:
The AI Project Cycle consists of the following stages:
1. Problem Scoping – Identifies the issue, stakeholders, and expected outcomes.
2. Data Acquisition – Collects relevant data for training and testing.
3. Data Exploration – Analyzes and visualizes data to find patterns.
4. Model Building – Trains an AI model using machine learning techniques.
Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 17 -
 5. Evaluation – Tests the model’s accuracy and effectiveness before deployment.
Explanation:
Each stage ensures that AI models are developed systematically and perform efficiently. Skipping
any stage can lead to poor predictions and unreliable results.

2. How does Natural Language Processing (NLP) help AI understand human language?
Answer:
NLP enables AI to process and understand text and speech by:
 Tokenization – Breaking text into words or phrases.
 Sentiment Analysis – Detecting emotions (positive, negative, neutral).
 Named Entity Recognition (NER) – Identifying names, places, and dates in text.
Explanation:
NLP powers applications like chatbots, voice assistants, and translation services, allowing AI to
interact with humans naturally and accurately.

3. What are some real-world applications of Data Science in the finance industry?
Answer:
Data Science is used in finance for:
1. Fraud Detection – Identifying unusual transactions using machine learning.
2. Risk Assessment – Evaluating loan applicants based on credit history.
3. Algorithmic Trading – Making stock market predictions using AI models.
Explanation:
By analyzing large datasets, financial institutions improve security, decision-making, and
efficiency, reducing risks and maximizing profits.

4. How does the K-Nearest Neighbors (KNN) algorithm work, and when is it used?
Answer:
KNN classifies data points by:
1. Measuring distance between a new data point and existing points (e.g., Euclidean distance).
2. Selecting K nearest neighbors and checking their labels.
3. Assigning the most common label to the new data point.
Explanation:
KNN is used in image recognition, recommendation systems, and medical diagnosis, where data
points with similar characteristics are grouped together.

5. What are the advantages of using Python in Data Science?

Answer:
Python is widely used in Data Science because:
1. Rich Libraries – Libraries like Pandas, NumPy, and Matplotlib simplify data handling.
2. Easy-to-Read Syntax – Reduces complexity, making it beginner-friendly.
3. Scalability & Community Support – Large open-source community helps with
troubleshooting and continuous improvement.
Explanation:
Python’s flexibility and efficiency make it ideal for data analysis, machine learning, and AI
applications, improving development speed and accuracy.

6. What is the difference between Precision, Recall, and F1-score in classification models?
Answer:
 Precision = (True Positives) / (True Positives + False Positives) → Measures accuracy of
positive predictions.
 Recall = (True Positives) / (True Positives + False Negatives) → Measures ability to detect
actual positives.
 F1-Score = 2 × (Precision × Recall) / (Precision + Recall) → Balances Precision and Recall.
Explanation:

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 18 -

These metrics help evaluate a classification model's effectiveness, especially in imbalanced
datasets (e.g., fraud detection where false negatives are costly).

7. How does a Decision Tree algorithm work in Machine Learning?

Answer:
A Decision Tree:
1. Splits data into branches based on feature conditions (e.g., "Is age > 30?").
2. Forms a tree-like structure, with leaves as final decisions.
3. Uses Gini Index or Entropy to decide the best split.
Explanation:
Decision Trees are used for classification and regression problems in areas like medical diagnosis
and customer segmentation due to their interpretability.

8. What are the key differences between SQL and NoSQL databases?
Answer:
Feature SQL Databases NoSQL Databases
Data Type Structured (Tables) Semi-structured/Unstructured (JSON, Graphs)
Scalability Vertical Scaling Horizontal Scaling
Use Case Financial Systems, ERPs Big Data, Real-time Analytics
Explanation:
SQL databases work well for structured transactional data, while NoSQL is better suited for
scalable, flexible data storage, such as social media applications.

9. What are the different types of Data Visualization techniques used in Data Science?
Answer:
1. Bar Charts – Compare categorical data.
2. Histograms – Show distribution of numerical data.
3. Scatter Plots – Visualize relationships between two variables.
4. Box Plots – Identify outliers and quartiles.
Explanation:
Data visualization tools like Matplotlib and Seaborn help in making data more understandable,
improving decision-making and insights extraction.

10. What is the role of clustering in Machine Learning, and how does K-Means Clustering
work?
Answer:
Role of Clustering:
 Groups similar data points without predefined labels.
 Helps in customer segmentation, anomaly detection, and document classification.
K-Means Clustering Steps:
1. Select K centroids (initial cluster centers).
2. Assign each data point to the nearest centroid.
3. Update centroids based on assigned points.
4. Repeat until clusters stabilize.
Explanation:
Clustering helps uncover hidden patterns in data, improving insights for marketing, cybersecurity,
and recommendation systems.

11. What are the different types of Machine Learning algorithms? Explain each briefly.
Answer:
1. Supervised Learning – The model is trained using labeled data. Examples: Linear
Regression, Decision Trees, Support Vector Machines (SVM).

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 19 -

 2. Unsupervised Learning – The model identifies patterns in unlabeled data. Examples: K-
Means Clustering, Principal Component Analysis (PCA).
3. Reinforcement Learning – The model learns by interacting with the environment and
receiving rewards or penalties. Examples: Q-Learning, Deep Q-Networks (DQN).
Explanation:
Machine Learning algorithms are categorized based on how they learn from data. Supervised models
need labeled datasets, unsupervised models find hidden structures, and reinforcement learning
improves decision-making over time.

12. What is overfitting in Machine Learning, and how can it be prevented?

Answer:
Overfitting occurs when a machine learning model learns the training data too well, including noise
and outliers, reducing its ability to generalize to new data.
Prevention Techniques:
1. Cross-validation – Splitting data into training and validation sets.
2. Regularization (L1/L2) – Adds penalty terms to the model to avoid complexity.
3. Pruning (for Decision Trees) – Removing unnecessary branches.
4. Increasing Training Data – More data helps improve model generalization.
Explanation:
Overfitting leads to high accuracy on training data but poor performance on unseen data. Preventing it
ensures the model makes reliable predictions in real-world scenarios.

13. How does Data Science improve recommendation systems?

Answer:
Recommendation systems analyze user behavior to suggest relevant items using:
1. Collaborative Filtering – Suggests items based on user preferences and similarities with
others.
2. Content-Based Filtering – Recommends items similar to those the user has interacted with.
3. Hybrid Models – Combine collaborative and content-based filtering for better accuracy.
Explanation:
Platforms like Netflix, Amazon, and Spotify use Data Science techniques to personalize
recommendations, increasing user engagement and sales.

14. What are the key differences between Regression and Classification models?
Answer:
Feature Regression Classification
Output Type Continuous numerical values Discrete categories
Examples Predicting house prices, stock prices Spam detection, medical diagnosis
Algorithms Linear Regression, Polynomial Regression Decision Trees, Random Forest, SVM
Explanation:
Regression models predict numeric outputs, while classification models categorize data into distinct
groups. The choice depends on the problem type.

15. What is Principal Component Analysis (PCA) and why is it used in Data Science?
Answer:
PCA is a dimensionality reduction technique used to transform high-dimensional data into a lower-
dimensional space while retaining important patterns.
Uses of PCA:
1. Reduces computational complexity in large datasets.
2. Improves model performance by removing redundant features.
3. Enhances visualization by reducing data to 2D/3D.
Explanation:

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 20 -

PCA is crucial when dealing with high-dimensional data, helping in feature selection and improving
machine learning model efficiency.

16. What are different evaluation metrics used for regression models?
Answer:
1. Mean Absolute Error (MAE) – Measures the average absolute difference between predicted
and actual values.
2. Mean Squared Error (MSE) – Calculates the average squared difference (gives more weight
to large errors).
3. R-Squared (R²) – Indicates how well the model explains the variance in the dataset (ranges
from 0 to 1).
Explanation:
These metrics help assess the accuracy and reliability of a regression model by comparing
predictions with actual outcomes.

17. What are the different types of clustering techniques in Unsupervised Learning?
Answer:
1. K-Means Clustering – Divides data into K clusters based on centroid distance.
2. Hierarchical Clustering – Creates a tree-like structure (dendrogram) to group data points.
3. DBSCAN (Density-Based Clustering) – Groups dense areas and identifies outliers.
Explanation:
Clustering is widely used in customer segmentation, anomaly detection, and document
classification, helping businesses find patterns in large datasets.

18. What is A/B Testing and how is it used in Data Science?

Answer:
A/B Testing is a statistical experiment where two versions (A and B) of a product, webpage, or
algorithm are compared to see which performs better.
Steps in A/B Testing:
1. Randomly split users into two groups (A & B).
2. Expose Group A to the original version and Group B to the new version.
3. Measure key performance indicators (KPIs) like conversion rates.
4. Use statistical analysis to determine the better version.
Explanation:
A/B Testing helps businesses optimize marketing strategies, website designs, and user
experiences based on data-driven decisions.

19. How does Data Science contribute to Sentiment Analysis?

Answer:
Sentiment Analysis classifies text as positive, negative, or neutral using:
1. Lexicon-Based Approaches – Uses a predefined dictionary of words with sentiment scores.
2. Machine Learning Models – Trains AI models on labeled datasets to recognize sentiment
patterns.
3. Deep Learning Techniques – Uses neural networks like LSTMs and Transformers for
advanced sentiment detection.
Explanation:
Sentiment Analysis is widely used in customer feedback analysis, brand reputation monitoring,
and social media trend analysis.

20. What is Time Series Analysis, and how is it used in forecasting?

Answer:
Time Series Analysis is a technique to analyze data points collected over time to identify trends,
seasonal effects, and patterns for forecasting.
Common Time Series Models:

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 21 -

 1. ARIMA (AutoRegressive Integrated Moving Average) – Predicts future values using past
trends.
2. LSTMs (Long Short-Term Memory Networks) – Deep learning model for sequential data
analysis.
3. Exponential Smoothing – Gives more weight to recent observations for accurate forecasting.
Explanation:
Time Series Analysis is used in stock market prediction, weather forecasting, and sales
forecasting, helping businesses make data-driven decisions.

21. What are the key differences between Structured, Semi-Structured, and Unstructured
Data?
Answer:
1. Structured Data – Organized in a predefined format (e.g., relational databases, spreadsheets).
2. Semi-Structured Data – Partially organized but not in a fixed format (e.g., JSON, XML
files).
3. Unstructured Data – No predefined structure (e.g., images, videos, social media posts).
Explanation:
Understanding data types is crucial in Data Science, as structured data is easy to analyze, while
unstructured data requires advanced AI techniques like Natural Language Processing (NLP) or
Computer Vision.

22. What is Data Wrangling, and why is it important in Data Science?

Answer:
Data Wrangling is the process of cleaning, transforming, and organizing raw data into a structured
format for analysis.
Key Steps:
1. Handling Missing Values – Filling or removing incomplete data points.
2. Removing Duplicates & Errors – Ensuring data consistency.
3. Transforming Data Formats – Converting data types (e.g., text to numerical).
Explanation:
Data Wrangling is essential as raw data is often messy and unstructured. Proper wrangling improves
data quality, leading to better model performance.

23. What is Cross-Validation, and how does it improve Machine Learning models?
Answer:
Cross-Validation is a technique to split the dataset into multiple training and testing sets to check a
model’s generalization ability.
Types of Cross-Validation:
1. K-Fold Cross-Validation – Divides data into K parts and iterates K times.
2. Leave-One-Out Cross-Validation (LOOCV) – Uses all but one data point for training and
tests on the remaining one.
Explanation:
Cross-validation helps detect overfitting, ensures robust model performance, and provides a more
reliable accuracy score for real-world data.

24. How does Decision Tree differ from Random Forest in Machine Learning?
Answer:
Feature Decision Tree Random Forest
Model Type Single tree-based model Ensemble of multiple decision trees
Accuracy Prone to overfitting More accurate and robust
Speed Faster for small datasets Slower due to multiple trees
Explanation:

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 22 -

Decision Trees are simple and interpretable, but prone to overfitting, while Random Forest
combines multiple trees for better accuracy and stability.

25. What are the different Distance Metrics used in Machine Learning?
Answer:
1. Euclidean Distance – Measures straight-line distance between two points.
2. Manhattan Distance – Measures distance along grid-based paths (like city blocks).
3. Cosine Similarity – Measures the angle between two vectors (used in NLP).
Explanation:
Distance metrics are crucial in clustering (K-Means), classification (KNN), and text similarity
analysis (NLP) for comparing data points effectively.

26. What is Feature Scaling, and what are its two main techniques?
Answer:
Feature Scaling standardizes data so that all variables contribute equally to the model.
Two main techniques:
1. Normalization (Min-Max Scaling) – Scales values between 0 and 1.
2. Standardization (Z-score Scaling) – Centers data around mean 0 with a standard deviation
of 1.
Explanation:
Feature Scaling is essential in algorithms like KNN, SVM, and Gradient Descent, where feature
magnitudes impact model performance.

27. What are Anomalies in Data Science, and how are they detected?
Answer:
Anomalies are data points that deviate significantly from expected patterns.
Anomaly Detection Techniques:
1. Statistical Methods – Use Z-score and Interquartile Range (IQR) to find outliers.
2. Machine Learning Models – Use Isolation Forests and Autoencoders for anomaly detection.
3. Rule-Based Methods – Set predefined thresholds for identifying anomalies.
Explanation:
Anomaly detection is widely used in fraud detection, cybersecurity, and quality control to identify
unusual events in datasets.

28. What are the advantages and limitations of using Neural Networks in AI?
Answer:
Advantages:
1. High accuracy in complex problems like image recognition and NLP.
2. Ability to learn non-linear relationships in large datasets.
3. Automatic feature extraction from raw data.
Limitations:
1. Requires large amounts of training data to perform well.
2. Computationally expensive due to multiple layers.
3. Lack of interpretability (black-box nature).
Explanation:
Neural Networks power deep learning models like CNNs for images and RNNs for sequences, but
they require significant resources and training time.

29. How is Clustering different from Classification in Machine Learning?

Answer:
Feature Clustering Classification
Learning Type Unsupervised Supervised
Labels Available? No (groups data automatically) Yes (predefined categories)

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 23 -

Feature Clustering Classification
Example K-Means, DBSCAN, Hierarchical Decision Trees, SVM, Logistic
Algorithms Clustering Regression
Explanation:
Clustering groups similar data points without predefined labels, while Classification assigns data
to known categories, making them suitable for different AI applications.

30. What is the purpose of an ROC Curve, and how is it used to evaluate models?
Answer:
An ROC (Receiver Operating Characteristic) Curve is a graphical plot that shows the trade-off
between True Positive Rate (Sensitivity) and False Positive Rate (1-Specificity) for a classification
model.
Key Concepts:
1. AUC (Area Under Curve) – Measures overall model performance (closer to 1 is better).
2. Higher ROC curve – Indicates a better-performing classifier.
3. Used in medical diagnostics, fraud detection, and risk assessment.
Explanation:
The ROC Curve helps compare models, showing their effectiveness in distinguishing between
different classes. A model with AUC = 0.5 performs no better than random guessing.

QUESTIONS AND ANSWERS - 5 marks

1. What are the key stages of the AI Project Cycle? Explain each stage in detail.
Answer:
The AI Project Cycle consists of five key stages that help in solving real-world problems using
Artificial Intelligence (AI).
1. Problem Scoping
o Identifies the issue that needs to be solved using AI.
o Defines stakeholders, project goals, and success criteria.
2. Data Acquisition
o Collects relevant data needed to train AI models.
o Data sources include online databases, government portals, and surveys.
3. Data Exploration
o Analyzes data for patterns, trends, and inconsistencies.
o Uses visualization tools like Matplotlib and Seaborn.
4. Model Building
o Selects appropriate Machine Learning models for predictions.
o Trains models using algorithms like Decision Trees, KNN, and Neural Networks.
5. Evaluation and Deployment
o Tests model performance using accuracy, precision, recall, and F1-score.
o Deploys AI models for real-world applications after fine-tuning.
Explanation:
Each stage ensures that AI models are developed systematically to deliver reliable and data-driven
solutions.

2. What are the different types of Machine Learning algorithms? Explain each with examples.
Answer:
Machine Learning algorithms are classified into three types:
1. Supervised Learning
o Trained on labeled data (input-output pairs).
o Example: Spam Detection using Naïve Bayes, where emails are classified as spam or
non-spam.
2. Unsupervised Learning
Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 24 -
 o Learns from unlabeled data and finds hidden patterns.
o Example: Customer Segmentation using K-Means Clustering, where similar
customers are grouped.
3. Reinforcement Learning
o Learns through trial and error using rewards and penalties.
o Example: Self-driving cars using Deep Q-Learning to improve navigation.
Explanation:
Each algorithm type serves different purposes, such as classification, clustering, and decision-
making, making them useful for real-world AI applications.

3. Explain the importance of Data Preprocessing in Data Science. What are its main steps?
Answer:
Data Preprocessing is a crucial step in Data Science that improves data quality and prepares it for
analysis.
Main Steps of Data Preprocessing:
1. Handling Missing Values – Filling gaps using mean, median, or mode.
2. Removing Duplicates – Eliminating repeated records to avoid redundancy.
3. Feature Scaling – Normalizing or standardizing numerical data.
4. Encoding Categorical Variables – Converting text data into numerical form using One-Hot
Encoding or Label Encoding.
5. Data Transformation – Converting raw data into a suitable format.
Explanation:
Data Preprocessing ensures that AI models work efficiently and accurately by removing
inconsistencies and improving dataset reliability.

4. What is K-Means Clustering? Explain its working with an example.

Answer:
K-Means Clustering is an unsupervised learning algorithm used to group similar data points into
clusters.
Working of K-Means:
1. Select the number of clusters (K).
2. Randomly initialize K cluster centroids.
3. Assign each data point to the nearest centroid.
4. Update centroids by calculating the mean of assigned points.
5. Repeat steps until centroids stabilize.
Example:
In customer segmentation, K-Means groups customers based on purchase behavior, allowing
businesses to target specific groups with personalized marketing.
Explanation:
K-Means is widely used in market segmentation, image compression, and anomaly detection to
uncover hidden patterns in data.

5. What are Neural Networks? Explain their structure and working.

Answer:
Neural Networks are AI models inspired by the human brain, used for deep learning applications.
Structure of a Neural Network:
1. Input Layer – Receives raw data.
2. Hidden Layers – Perform computations using weighted connections.
3. Output Layer – Provides final predictions.
Working of a Neural Network:
 Each neuron applies an activation function (ReLU, Sigmoid, etc.).
 Weights and biases are adjusted using Backpropagation and Gradient Descent.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 25 -

Example:
Neural Networks power applications like image recognition, speech processing, and language
translation.
Explanation:
Deep Neural Networks improve AI capabilities, enabling tasks like autonomous driving, medical
diagnostics, and fraud detection.

6. What is Time Series Analysis? How is it used in forecasting?

Answer:
Time Series Analysis is a method used to analyze and predict future values based on historical data
trends.
Steps in Time Series Analysis:
1. Collect time-stamped data (e.g., stock prices, weather reports).
2. Identify trends and seasonality in data.
3. Apply forecasting models like ARIMA, LSTMs, or Exponential Smoothing.
Example:
Time Series Analysis is used for sales forecasting, where businesses predict future sales based on
past purchase trends.
Explanation:
Accurate time series forecasting helps industries optimize inventory, manage risks, and improve
decision-making.

7. What is Principal Component Analysis (PCA), and why is it used in Data Science?
Answer:
PCA is a dimensionality reduction technique used to simplify complex datasets.
Why PCA is Used:
1. Reduces computational cost in Machine Learning.
2. Removes redundancy by transforming correlated variables.
3. Improves model accuracy by reducing noise.
Example:
In facial recognition systems, PCA helps extract essential features while ignoring unnecessary
details.
Explanation:
PCA improves efficiency and visualization in large datasets, making it essential for high-
dimensional data analysis.

8. Explain the difference between Precision, Recall, and F1-score in classification models.
Answer:
 Precision – Measures the accuracy of positive predictions. Formula:

 Recall – Measures the ability to detect all positive instances. Formula:

 F1-score – Harmonic mean of Precision and Recall. Formula:

Explanation:
These metrics evaluate classification performance, especially in imbalanced datasets like fraud
detection and medical diagnosis.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 26 -

9. What are the advantages and challenges of using Big Data in Data Science?
Answer:
Advantages:
1. Helps in better decision-making using AI insights.
2. Improves customer personalization in businesses.
3. Enables real-time monitoring (e.g., IoT applications).
Challenges:
1. High storage costs for massive datasets.
2. Data security and privacy risks.
3. Complexity in processing large-scale data.
Explanation:
Big Data is essential for industries like healthcare, finance, and e-commerce, but it requires
advanced tools for efficient handling.

10. How does Sentiment Analysis work in Natural Language Processing (NLP)?
Answer:
Sentiment Analysis classifies text as positive, negative, or neutral using NLP techniques.
Steps in Sentiment Analysis:
1. Tokenization – Breaking text into words.
2. Removing Stopwords – Filtering unnecessary words.
3. Applying Machine Learning – Using models like Naïve Bayes or Transformer Networks.
Example:
Social media platforms analyze tweets to determine public opinion on trending topics.
Explanation:
Sentiment Analysis helps brands monitor customer feedback, brand reputation, and political
sentiment analysis.

11. What are the different types of Data and their significance in Data Science?
Answer:
Data in Data Science is categorized into three main types:
1. Structured Data
o Organized in rows and columns (e.g., relational databases, spreadsheets).
o Example: Bank transaction records.
2. Semi-Structured Data
o Partially organized data with some structure but not in a tabular format.
o Example: JSON and XML files used in web applications.
3. Unstructured Data
o No predefined format, making it difficult to store in relational databases.
o Example: Social media posts, images, videos, audio files.
Explanation:
Understanding these data types helps Data Scientists select appropriate storage methods,
processing techniques, and analytical models for different use cases.

12. Explain the importance of Exploratory Data Analysis (EDA) in Data Science. What are its
key steps?
Answer:
EDA is the process of analyzing datasets to summarize their key characteristics before applying
Machine Learning models.
Key Steps in EDA:
1. Understanding Data Types – Identifying numerical, categorical, and textual data.
2. Handling Missing Values – Filling gaps using mean, median, or mode.
3. Outlier Detection – Identifying unusual values using Box Plots and Z-scores.
4. Feature Correlation Analysis – Checking relationships between variables using heatmaps.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 27 -

 5. Data Visualization – Using histograms, scatter plots, and bar charts to identify trends.
Explanation:
EDA helps in cleaning, transforming, and understanding data, improving the accuracy and
efficiency of Machine Learning models.

13. What is Feature Engineering? Discuss different techniques used for Feature Engineering.
Answer:
Feature Engineering is the process of creating new features from existing data to improve Machine
Learning model performance.
Techniques of Feature Engineering:
1. Feature Scaling – Normalizing numerical values using Min-Max Scaling or Standardization.
2. Feature Extraction – Creating new features from raw data (e.g., extracting text length from
customer reviews).
3. Feature Encoding – Converting categorical variables into numerical form using One-Hot
Encoding or Label Encoding.
4. Feature Selection – Identifying the most relevant features using Principal Component
Analysis (PCA) or Recursive Feature Elimination (RFE).
Explanation:
Feature Engineering helps enhance model accuracy by providing better input data for training,
making it a critical step in Data Science.

14. What are the different types of Regression models in Machine Learning? Explain with
examples.
Answer:
Regression models are used to predict continuous numerical values based on input variables.
Types of Regression Models:
1. Linear Regression – Models relationships between dependent and independent variables
using a straight line.
o Example: Predicting house prices based on area size.
2. Polynomial Regression – Extends Linear Regression by adding polynomial terms.
o Example: Predicting population growth trends.
3. Logistic Regression – Used for binary classification problems.
o Example: Predicting whether an email is spam or not.
4. Ridge and Lasso Regression – Add regularization to prevent overfitting.
o Example: Reducing complexity in stock price prediction models.
Explanation:
Choosing the right regression model depends on data patterns, complexity, and the number of
independent variables.

15. What is Big Data? Explain the 5Vs of Big Data with examples.
Answer:
Big Data refers to large and complex datasets that cannot be processed using traditional methods.
The 5Vs of Big Data:
1. Volume – The amount of data generated (e.g., Google processes 20 petabytes of data daily).
2. Velocity – The speed at which data is created and processed (e.g., real-time social media
feeds).
3. Variety – Different types of data (structured, semi-structured, unstructured).
4. Veracity – The accuracy and reliability of data (e.g., financial transaction records must be
error-free).
5. Value – The usefulness of data for decision-making (e.g., customer behavior analysis for
targeted marketing).
Explanation:
Big Data enables organizations to make data-driven decisions but requires advanced technologies
like Hadoop and Spark for processing.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 28 -

16. Explain Decision Trees in Machine Learning. How do they work?
Answer:
Decision Trees are a type of Supervised Learning algorithm used for classification and regression.
How Decision Trees Work:
1. Root Node – The starting point of the tree representing the entire dataset.
2. Splitting – The dataset is split based on a feature using Gini Index or Entropy.
3. Decision Nodes – Intermediate nodes where further splitting occurs.
4. Leaf Nodes – The final classification or prediction result.
Example:
A Decision Tree can be used to classify whether a customer will buy a product based on income and
previous purchase history.
Explanation:
Decision Trees are easy to interpret but prone to overfitting, which is mitigated using Random
Forests.

17. What are Recurrent Neural Networks (RNNs)? How are they used in AI applications?
Answer:
RNNs are a type of Deep Learning model designed for sequential data processing.
How RNNs Work:
1. Maintain a memory of previous inputs using recurrent connections.
2. Use Hidden States to process sequences.
3. Apply Backpropagation Through Time (BPTT) to update weights.
Applications of RNNs:
1. Speech Recognition – AI assistants like Siri and Google Assistant.
2. Machine Translation – Google Translate for language conversion.
3. Stock Market Prediction – Analyzing historical trends to forecast prices.
Explanation:
RNNs excel at time-dependent tasks but suffer from the vanishing gradient problem, which is
improved by LSTMs and GRUs.

18. What is Model Overfitting? How can it be prevented?

Answer:
Overfitting occurs when a Machine Learning model learns the training data too well, including
noise, leading to poor generalization.
Techniques to Prevent Overfitting:
1. Cross-Validation – Splitting data into training and testing sets multiple times.
2. Regularization (L1 & L2) – Adding penalty terms to the model.
3. Dropout in Neural Networks – Randomly deactivating neurons to prevent reliance on
specific nodes.
4. Increasing Training Data – More diverse data helps in better learning.
Explanation:
Preventing overfitting ensures that the model performs well on unseen data, making it more reliable
for real-world applications.

19. What is Transfer Learning? How is it used in Deep Learning?

Answer:
Transfer Learning is an AI technique where a pre-trained model is adapted for a new but related task.
How Transfer Learning Works:
1. A model is pre-trained on a large dataset (e.g., ImageNet for image classification).
2. The learned features are transferred to a new model for a specific task.
3. Only fine-tuning is required on the new dataset.
Examples:
 Computer Vision – Using pre-trained models like ResNet for medical image analysis.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 29 -

  NLP – GPT and BERT models for text summarization and translation.
Explanation:
Transfer Learning reduces computation time, improves accuracy, and allows training on small
datasets.

20. How does AI impact industries like Healthcare, Finance, and Education?
Answer:
1. Healthcare – AI is used for disease prediction, robotic surgeries, and drug discovery.
2. Finance – AI detects fraud, automates trading, and provides personalized financial
advice.
3. Education – AI enables personalized learning, automated grading, and AI tutors.
Explanation:
AI revolutionizes multiple industries by enhancing efficiency, reducing costs, and improving
decision-making, making it one of the most influential technologies today.

Prepared by: M. S. KumarSwamy, TGT(Maths) Page - 30 -