QUESTIONS/ANSWERS OF MACHINE
LEARNING
1. Explain what is the function of ‘Unsupervised Learning’?
Ans : Unsupervised learning is a type of machine learning where algorithms learn
patterns from unlabeled data without any human intervention. It's like teaching a
1
child to recognize different shapes by showing them various objects without labeling
them. 2
Key Function:
Discover Hidden Patterns: Unsupervised learning algorithms analyze large
datasets to identify underlying structures, relationships, and anomalies that might not
be apparent to humans. 3
Group Similar Data Points: It automatically clusters data points with similar
characteristics, enabling us to categorize and segment data effectively. 4
Reduce Data Dimensionality: It helps simplify complex datasets by reducing the
number of features while preserving essential information. 5
Anomaly Detection: It can identify outliers or unusual data points that might indicate
errors, fraud, or other important insights. 6
Common Use Cases:
Customer Segmentation: Grouping customers based on their behaviors and
preferences to tailor marketing strategies. 7
Image and Document Clustering: Organizing large collections of images or
documents into meaningful categories. 8
Feature Extraction: Identifying the most relevant features from a dataset to improve
the performance of other machine learning models. 9
Anomaly Detection in Network Traffic: Detecting unusual network activity that
could signal security threats.
Key Algorithms:
Clustering Algorithms:
o K-Means Clustering 10
o Hierarchical Clustering 11
o DBSCAN 12
Dimensionality Reduction Techniques:
o Principal Component Analysis (PCA) 13
,o t-SNE (t-Distributed Stochastic Neighbor Embedding) 14
Advantages:
No Need for Labeled Data: Saves time and effort in data preparation. 15
Discover Unexpected Patterns: Can reveal insights that might be missed with human
analysis. 16
Versatile: Applicable to a wide range of data types and domains. 17
Limitations:
Interpretation Challenges: Understanding the meaning of discovered patterns can be
complex.
Sensitivity to Initial Conditions: Some algorithms can produce different results
depending on the starting point.
In conclusion, unsupervised learning is a powerful tool for exploring and
understanding large datasets, making it a valuable asset in various fields, including
data science, business analytics, and artificial intelligence. 18
2. What do you understand by Eigenvectors and Eigenvalues?
Ans : Eigenvectors and Eigenvalues in Machine Learning
In the realm of machine learning, eigenvectors and eigenvalues are fundamental
concepts borrowed from linear algebra. They offer valuable insights into the
1
structure and properties of data, enabling us to perform tasks like dimensionality
reduction, feature extraction, and more. 2
What are Eigenvectors and Eigenvalues?
Eigenvector: An eigenvector of a square matrix is a non-zero vector that, when
multiplied by the matrix, results in a scalar multiple of itself. In other words, it's a
3 4
direction that remains unchanged after a linear transformation, though its magnitude
may be scaled. 5
Eigenvalue: The corresponding scalar factor by which the eigenvector is scaled is
called the eigenvalue. It represents the strength or impact of the transformation
6
along the direction of the eigenvector. 7
Why are They Important in Machine Learning?
1. Dimensionality Reduction:
o Principal Component Analysis (PCA): PCA leverages eigenvectors to identify the
principal components of a dataset, which are the directions of maximum variance. 8
By projecting data onto these principal components, we can reduce dimensionality
while preserving most of the information. 9
,2. Feature Extraction:
o Feature Engineering: Eigenvectors can help identify the most informative features in
a dataset. By selecting eigenvectors associated with the largest eigenvalues, we
10
can extract the most significant features. 11
3. Image and Signal Processing:
o Image Compression: Eigenvectors can be used to decompose images into their
principal components, allowing for efficient compression. 12
o Signal Denoising: Eigenvectors can help identify and remove noise from signals. 13
4. Clustering and Classification:
o Clustering: Eigenvectors can be used to find underlying patterns in data, aiding in
clustering algorithms. 14
o Classification: Eigenvectors can be used to extract relevant features for
classification models. 15
Key Points to Remember:
Eigenvectors and eigenvalues are mathematical concepts that provide insights into
the structure of data. 16
They are used in various machine learning techniques, including dimensionality
reduction, feature extraction, and image processing. 17
By understanding eigenvectors and eigenvalues, we can gain deeper insights into
data and build more effective machine learning models. 18
In essence, eigenvectors and eigenvalues are powerful tools that enable us to
simplify complex data, extract meaningful information, and make better-informed
decisions in the field of machine learning. 19
3.List different forms of learning.
Ans: There are typically five main forms of learning in Machine Learning (ML):
1. Supervised Learning: This is where the algorithm learns from labeled data. The
data has both input features and desired output values (labels). The goal is to train
the model to map inputs to the correct outputs, enabling it to make predictions or
classifications for new, unseen data.
o Examples: Spam filtering, image recognition, weather forecasting.
o Common Algorithms: Linear Regression, Logistic Regression, Support Vector
Machines, Decision Trees.
, 2. Unsupervised Learning: This involves learning patterns from unlabeled data, where
data points have no predefined labels or categories. The model identifies hidden
structures, relationships, and anomalies in the data.
o Examples: Customer segmentation, image/document clustering, anomaly detection
in network traffic.
o Common Algorithms: K-Means Clustering, Hierarchical Clustering, Principal
Component Analysis (PCA).
3. Semi-Supervised Learning: This combines labeled and unlabeled data for training.
1
It leverages the power of labeled data while utilizing the abundance of unlabeled
data to improve model performance.
o Examples: Sentiment analysis with limited labeled examples, image classification
with a small set of labeled images.
4. Self-Supervised Learning: This utilizes unlabeled data by creating artificial labels
or tasks for the model to learn from. It allows the model to learn generalizable
representations of the data without explicit human supervision.
o Examples: Pre-training language models on large text corpora, image feature
extraction using self-supervised tasks.
5. Reinforcement Learning: This involves training a model through trial and error in a
simulated environment. The model receives rewards or penalties for its actions and
learns to optimize its behavior to maximize rewards.
o Examples: Training a robot to navigate a maze, training an AI agent to play games.
These learning forms offer different approaches to train models, making machine
learning versatile and applicable to various tasks and data types.
4.Identify the disadvantage of K- NN algorithm.
Ans: Disadvantages of the K-NN Algorithm
While the K-NN algorithm is simple and intuitive, it has several drawbacks:
1. Computational Cost:
o High Memory Usage: K-NN requires storing the entire training dataset, which can be
memory-intensive for large datasets. 1
o Slow Prediction: During the prediction phase, the algorithm calculates distances to
all training points, making it computationally expensive, especially for large datasets. 2
2. Sensitivity to the Choice of K:
LEARNING
1. Explain what is the function of ‘Unsupervised Learning’?
Ans : Unsupervised learning is a type of machine learning where algorithms learn
patterns from unlabeled data without any human intervention. It's like teaching a
1
child to recognize different shapes by showing them various objects without labeling
them. 2
Key Function:
Discover Hidden Patterns: Unsupervised learning algorithms analyze large
datasets to identify underlying structures, relationships, and anomalies that might not
be apparent to humans. 3
Group Similar Data Points: It automatically clusters data points with similar
characteristics, enabling us to categorize and segment data effectively. 4
Reduce Data Dimensionality: It helps simplify complex datasets by reducing the
number of features while preserving essential information. 5
Anomaly Detection: It can identify outliers or unusual data points that might indicate
errors, fraud, or other important insights. 6
Common Use Cases:
Customer Segmentation: Grouping customers based on their behaviors and
preferences to tailor marketing strategies. 7
Image and Document Clustering: Organizing large collections of images or
documents into meaningful categories. 8
Feature Extraction: Identifying the most relevant features from a dataset to improve
the performance of other machine learning models. 9
Anomaly Detection in Network Traffic: Detecting unusual network activity that
could signal security threats.
Key Algorithms:
Clustering Algorithms:
o K-Means Clustering 10
o Hierarchical Clustering 11
o DBSCAN 12
Dimensionality Reduction Techniques:
o Principal Component Analysis (PCA) 13
,o t-SNE (t-Distributed Stochastic Neighbor Embedding) 14
Advantages:
No Need for Labeled Data: Saves time and effort in data preparation. 15
Discover Unexpected Patterns: Can reveal insights that might be missed with human
analysis. 16
Versatile: Applicable to a wide range of data types and domains. 17
Limitations:
Interpretation Challenges: Understanding the meaning of discovered patterns can be
complex.
Sensitivity to Initial Conditions: Some algorithms can produce different results
depending on the starting point.
In conclusion, unsupervised learning is a powerful tool for exploring and
understanding large datasets, making it a valuable asset in various fields, including
data science, business analytics, and artificial intelligence. 18
2. What do you understand by Eigenvectors and Eigenvalues?
Ans : Eigenvectors and Eigenvalues in Machine Learning
In the realm of machine learning, eigenvectors and eigenvalues are fundamental
concepts borrowed from linear algebra. They offer valuable insights into the
1
structure and properties of data, enabling us to perform tasks like dimensionality
reduction, feature extraction, and more. 2
What are Eigenvectors and Eigenvalues?
Eigenvector: An eigenvector of a square matrix is a non-zero vector that, when
multiplied by the matrix, results in a scalar multiple of itself. In other words, it's a
3 4
direction that remains unchanged after a linear transformation, though its magnitude
may be scaled. 5
Eigenvalue: The corresponding scalar factor by which the eigenvector is scaled is
called the eigenvalue. It represents the strength or impact of the transformation
6
along the direction of the eigenvector. 7
Why are They Important in Machine Learning?
1. Dimensionality Reduction:
o Principal Component Analysis (PCA): PCA leverages eigenvectors to identify the
principal components of a dataset, which are the directions of maximum variance. 8
By projecting data onto these principal components, we can reduce dimensionality
while preserving most of the information. 9
,2. Feature Extraction:
o Feature Engineering: Eigenvectors can help identify the most informative features in
a dataset. By selecting eigenvectors associated with the largest eigenvalues, we
10
can extract the most significant features. 11
3. Image and Signal Processing:
o Image Compression: Eigenvectors can be used to decompose images into their
principal components, allowing for efficient compression. 12
o Signal Denoising: Eigenvectors can help identify and remove noise from signals. 13
4. Clustering and Classification:
o Clustering: Eigenvectors can be used to find underlying patterns in data, aiding in
clustering algorithms. 14
o Classification: Eigenvectors can be used to extract relevant features for
classification models. 15
Key Points to Remember:
Eigenvectors and eigenvalues are mathematical concepts that provide insights into
the structure of data. 16
They are used in various machine learning techniques, including dimensionality
reduction, feature extraction, and image processing. 17
By understanding eigenvectors and eigenvalues, we can gain deeper insights into
data and build more effective machine learning models. 18
In essence, eigenvectors and eigenvalues are powerful tools that enable us to
simplify complex data, extract meaningful information, and make better-informed
decisions in the field of machine learning. 19
3.List different forms of learning.
Ans: There are typically five main forms of learning in Machine Learning (ML):
1. Supervised Learning: This is where the algorithm learns from labeled data. The
data has both input features and desired output values (labels). The goal is to train
the model to map inputs to the correct outputs, enabling it to make predictions or
classifications for new, unseen data.
o Examples: Spam filtering, image recognition, weather forecasting.
o Common Algorithms: Linear Regression, Logistic Regression, Support Vector
Machines, Decision Trees.
, 2. Unsupervised Learning: This involves learning patterns from unlabeled data, where
data points have no predefined labels or categories. The model identifies hidden
structures, relationships, and anomalies in the data.
o Examples: Customer segmentation, image/document clustering, anomaly detection
in network traffic.
o Common Algorithms: K-Means Clustering, Hierarchical Clustering, Principal
Component Analysis (PCA).
3. Semi-Supervised Learning: This combines labeled and unlabeled data for training.
1
It leverages the power of labeled data while utilizing the abundance of unlabeled
data to improve model performance.
o Examples: Sentiment analysis with limited labeled examples, image classification
with a small set of labeled images.
4. Self-Supervised Learning: This utilizes unlabeled data by creating artificial labels
or tasks for the model to learn from. It allows the model to learn generalizable
representations of the data without explicit human supervision.
o Examples: Pre-training language models on large text corpora, image feature
extraction using self-supervised tasks.
5. Reinforcement Learning: This involves training a model through trial and error in a
simulated environment. The model receives rewards or penalties for its actions and
learns to optimize its behavior to maximize rewards.
o Examples: Training a robot to navigate a maze, training an AI agent to play games.
These learning forms offer different approaches to train models, making machine
learning versatile and applicable to various tasks and data types.
4.Identify the disadvantage of K- NN algorithm.
Ans: Disadvantages of the K-NN Algorithm
While the K-NN algorithm is simple and intuitive, it has several drawbacks:
1. Computational Cost:
o High Memory Usage: K-NN requires storing the entire training dataset, which can be
memory-intensive for large datasets. 1
o Slow Prediction: During the prediction phase, the algorithm calculates distances to
all training points, making it computationally expensive, especially for large datasets. 2
2. Sensitivity to the Choice of K:
