2. Self Introduction
Name: Dr. Chetan Jalendra
Qualification: Ph.D. (BITS Pilani)
Designation: Assistant Professor
Research Area: Robotics, AI & ML
Email: chetan.jalendra@jietjodhpur.ac.in
LinkedIn: linkedin.com/in/chetan-jalendra
Instructional Language: 80% English, 20% Hindi
3. Course Introduction
π Overview:
β’This course is designed to bridge foundational theories of machine learning
with real-world applications.
β’Emphasis is placed on hands-on learning, ensuring students can apply
concepts to real datasets and practical problems.
π Key Concepts Covered:
β’Feature Extraction: Linear Discriminant Analysis (LDA), Principal
Component Analysis (PCA), Singular Value Decomposition (SVD)
β’Feature Selection: Filter, Wrapper, and Embedded methods
β’Model Evaluation: Cross-validation, resampling, statistical hypothesis
testing, algorithm comparison
4. Course Introduction
β’Advanced Topics:
β’ Reinforcement Learning (Q-learning, SARSA, Policy Iteration)
β’ Semi-supervised Learning
β’ Recommender Systems (Collaborative & Content-based filtering)
π― Learning Outcomes:
β’Gain expertise in building, tuning, and evaluating machine learning models
β’Understand both theoretical and mathematical underpinnings of key
algorithms
β’Apply learned concepts to domains such as healthcare, finance, and
technology
β’Prepare for industry roles in Data Science, Machine Learning, AI, and academic
research
5. Machine Learning vs. Applied
Machine Learning
Aspect Machine Learning (ML) Applied Machine Learning (AML)
Focus
Theoretical understanding of
learning algorithms
Practical implementation of ML
techniques
Approach
Emphasizes mathematical
models, proofs, and algorithm
design
Emphasizes real-world problem
solving using data and tools
Tools Used Custom algorithms, research-
focused environments
Libraries like scikit-learn,
TensorFlow, pandas, etc.
Learning
Objective
Understand "how" and "why"
models work
Understand "how to use" models
effectively on real-world data
Applications
Research, developing new
algorithms
Industry projects, decision
systems, product development
Target
Audience
Researchers, theorists
Data scientists, analysts,
engineers
6. Relevance of the Course
Technical Domain
β’Applied Machine Learning is a core area in Computer Science, Artificial
Intelligence, Data Science, and Robotics.
β’It provides students with essential skills in:
β’ Data preprocessing and feature engineering
β’ Model training, tuning, and evaluation
β’ Deployment of predictive systems
β’Strengthens concepts from foundational subjects like programming,
statistics, and linear algebra.
β’Helps students understand and implement algorithms that solve real
engineering problems.
7. Relevance of the Course
Student Perspective
β’Enhances problem-solving, critical thinking, and analytical abilities.
β’Builds a strong portfolio for:
β’ Higher studies in AI/ML
β’ Competitive internships and job placements in top tech companies
β’Prepares students for interdisciplinary research and innovation.
β’Encourages lifelong learning in a rapidly evolving tech landscape.
8. Relevance of the Course
Relevance to Industry Needs
β’Machine learning is one of the most in-demand skills globally.
β’Companies are actively seeking graduates with hands-on experience in:
β’ Data modeling and prediction
β’ Business intelligence and analytics
β’ AI-based product development
β’Aligns with roles like ML Engineer, Data Scientist, AI Analyst, and
Business Intelligence Developer.
9. Unit I: Statistical Learning
Theory
β’Introduces the theoretical foundation of machine learning and model
generalization.
β’Covers feature extraction techniques like:
β’ Linear Discriminant Analysis (LDA) β class-separability
β’ Principal Component Analysis (PCA) β dimensionality reduction
β’ Singular Value Decomposition (SVD) β matrix factorization
β’Explains feature selection methods:
β’ Filter, Wrapper, and Embedded techniques
β’Helps in improving model performance by selecting relevant features
and reducing overfitting.
β’Essential for understanding data representation before model building.
10. Unit II: Analysis of Machine
Learning
β’Focuses on evaluating and comparing machine learning models
effectively.
β’Covers cross-validation and resampling methods to estimate model
performance.
β’Discusses classifier performance metrics: accuracy, precision, recall, F1-
score, etc.
β’Introduces hypothesis testing for validating model comparisons.
β’Explains how to compare two algorithms statistically using McNemar
test and others.
β’Covers analysis of variance (ANOVA) for comparing multiple models.
β’Applies evaluation techniques across different datasets for robust
analysis.
11. Unit III: Model Evaluation &
Algorithm Selection
β’Deep dive into validation techniques: holdout, k-fold cross-validation, nested
cross-validation.
β’Covers model evaluation metrics and uncertainty estimation using
bootstrapping.
β’Introduces hyperparameter optimization methods, including grid/random
search.
β’Explains algorithm comparison using statistical tests: McNemar Test, F-test,
and Proportion Testing.
β’Focus on selecting the best model/algorithm based on performance and
complexity.
β’Includes evolutionary approaches to optimize deep learning
hyperparameters.
12. Unit IV: Reinforcement
Learning
β’Focuses on learning from interactions with an environment to maximize
cumulative rewards.
β’Introduces Markov Decision Processes (MDP) and Bellman Equations for
modeling decision-making.
β’Covers Monte Carlo methods for policy evaluation.
β’Explains Policy Iteration and Value Iteration techniques.
β’Discusses model-free learning methods: Q-Learning and SARSA.
β’Introduces Model-based Reinforcement Learning for environments with
known dynamics.
β’Applicable in robotics, gaming, and autonomous systems.
13. Unit V: Semi-supervised Learning &
Recommender Systems
β’Semi-supervised Learning leverages both labeled and unlabeled data to
improve model performance.
β’Covers key aspects like confidence-based learning and self-training.
β’Useful in domains with limited labeled data, such as medical imaging or
speech recognition.
β’Recommender Systems help personalize content using user preferences
and behavior.
β’Covers Collaborative Filtering (user-item interactions) and Content-
based Filtering (item features).
β’Explores explicit (ratings) and implicit (clicks, views) feedback
mechanisms.
β’Widely used in e-commerce, streaming platforms, and social media.
14. Text & Reference Books
π Textbooks
T1: Introduction to Machine Learning with Python
Authors: Andreas C. MΓΌller, Sarah Guido β O'Reilly Media
T2: Ensemble Methods: Foundations and Algorithms
Author: Zhi-Hua Zhou β CRC Press
T3: Artificial Intelligence and Machine Learning
Author: Sudhansh Sharma β IGNOU
π Reference Books
R1: Applied Machine Learning
Author: David Forsyth β Springer
R2: Understanding Machine Learning: From Theory to Algorithms
Authors: Shai Shalev-Shwartz, Shai Ben-David β Cambridge University Press
R3: Neural Networks and Machine Learning
Author: Simon Haykin
15. Websites & Journals
π Recommended Websites
https://scikit-learn.org/ β Official documentation and tutorials
https://machinelearningmastery.com/ β Practical ML guides and examples
π Key Journals & Handbooks
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
IEEE Transactions on Neural Networks and Learning Systems
Elsevier Artificial Intelligence
Elsevier Neural Networks
IEEE Transactions on Knowledge and Data Engineering
Elsevier Applied Soft Computing
Elsevier Knowledge-Based Systems
16. Examination Scheme
1st Continuous Evaluation: 12 Marks β 30%
1st Midterm: 12 Marks β 60%
2nd Continuous Evaluation: 6 Marks β 40%
End Term Exam: 70 Marks β 100% Syllabus
17. Thank You
Dr. Chetan Jalendra
Assistant Professor I
Applied Machine Learning β Zero Lecture
