Details of CS5103 (Autumn 2025)

Level: 5 Type: Theory Credits: 4.0

Course CodeCourse NameInstructor(s)
CS5103 Applied Machine Learning Kripabandhu Ghosh,
Saptarshi Pyne

Syllabus
Learning to use Python machine learning libraries
o Writing and importing code in Python
o Learning to use the scikit-learn and TensorFlow libraries
Conducting classification tasks
o A brief review of classification algorithms or classifiers
o Recognizing hand-written digits using a support vector classifier
o Conducting a comparative study of the widely used classifiers using insilico datasets
o Investigating the class boundaries learned by the linear discriminant analysis (LDA) versus that of the quadratic discriminant analysis (QDA)
o Predicting whether the share market will move up or down using a naive Bayes classifier and a k-nearest neighbours (KNN) classifier
o Identifying the peak hours of bike rentals using linear regression and Poisson regression
Performing feature extraction
o Extracting features from text documents
o Extracting features from image files

Selecting linear models (with and without regularization) that best describe a
given dataset
o A brief review of linear model selection algorithms
o Choosing the models that best describe how salaries of baseball players vary
based on various statistics associated with the players performance: Using
least squares regression, partial least squares (PLS) regression, ridge
regression, lasso regression, and principal components regression
Utilizing tree-based algorithms for predication
o A brief review of tree-based algorithms
o Predicting housing prices with different types of decision trees
(classification trees, regression trees) and random forests (with and
without bagging and boosting)

Employing support vector machines (SVMs) for classification tasks
o A brief review of SVMs
o Employing an SVM for one-class classification
o Employing an SVM for two-class classification
o Employing an SVM for multi-class classification

Clustering
o A brief review of clustering algorithms
o Subgrouping cancer cell lines into cancer types using the k-means and
hierarchical clustering algorithms

Training and testing artificial neural networks (ANNs)
o A brief review of ANNs
o Deploying a single-layer ANN
o Deploying a multi-layer ANN

Modelling real-world systems with graphical models
o A brief review of graphical models
o Modelling a stock market using Bayesian networks
o Denoising images using Markov random fields (MRFs)
o Understanding speech with hidden Markov models (HMMs)

Demonstrating a few state-of-the-art applications of machine learning
o Application of large language models for legal text summarization
o Application of word embedding for information retrieval
o Classification of aerial photographs under scarcity of labelled samples
o Prediction of 3D protein structures with AlphaFold

References
Textbook
James, G., Witten, D., Hastie, T., Tibshirani, R., & Taylor, J. (2023). An Introduction
to Statistical Learning: With Applications in Python. Springer Nature.

Reference books
Bishop, C. M. (2006). Pattern Recognition and Machine Learning. Springer.
Duda, R. O., & Hart, P. E. (2006). Pattern Classification. John Wiley & Sons.
Marsland, S. (2011). Machine Learning: An Algorithmic Perspective. Chapman and Hall/CRC.

Course Credit Options

Sl. No.ProgrammeSemester NoCourse Choice
1 IP 1 Not Allowed
2 IP 3 Not Allowed
3 IP 5 Not Allowed
4 MP 1 Not Allowed
5 MP 3 Not Allowed
6 MR 1 Not Allowed
7 MR 3 Not Allowed
8 MS 3 Not Allowed
9 MS 5 Not Allowed
10 MS 7 Not Allowed
11 MS 9 Elective
12 RS 1 Not Allowed
13 RS 2 Not Allowed