Course Code Course name L T P C

CSAI7019 Applied Machine Learning 3 0 1 4

Total Units to be Covered: Total Contact Hours: 75
Syllabus version:
Prerequisite(s):
1.0

Course Objectives
1. Understand the core concepts and techniques of machine learning and artificial
intelligence.
2. Develop machine learning models using popular libraries and frameworks.
3. Evaluate the performance of machine learning models using appropriate metrics.
4. Apply machine learning to various real-world problems and domains.

Course Outcomes

On completion of this course, the students will be able to

CO1. Recall and define key machine learning concepts, terminologies, and
algorithms.
CO2. Describe the differences between supervised, unsupervised, and
reinforcement learning.
CO3. Apply data preprocessing techniques to clean, transform, and prepare
datasets for machine learning.
CO4. Apply, compare, and contrast the strengths and weaknesses of different
machine learning algorithms.

CO-PO Mapping

Progr
am
Outco
P P P P P P P P P
mes PO PO PO PS PS PS
O O O O O O O O O
Cours 10 11 12 O1 O2 O3
1 2 3 4 5 6 7 8 9
e
Outco
mes
CO 1 1 1 1 - - - - - - - - - 1 - 3
-
CO 2 1 1 1 - - - - - - - - - 1 3

CO 3 1 1 1 - - - - - - - - - 1 - 3
 -
CO4 1 1 1 - - - - - - - - - 2 3
-
Avera 1.2
1 1 1 - - - - - - - - - 3
ge 5

1 – Weakly Mapped (Low) 2 – Moderately Mapped (Medium)

3 – Strongly Mapped (High) “_” means there is no correlation

Syllabus

Unit I: Introduction 3 Lecture Hours

Overview of machine learning and its applications, Types of machine learning:
supervised, unsupervised, reinforcement, Python and libraries for machine learning
(e.g., NumPy, Pandas, scikit-learn)
Unit II: Loss functions 3 Lecture
Hours
Mean Squared Error (MSE), Mean Absolute Error (MAE), Huber Loss, Binary Cross-
Entropy Loss (Log Loss), Categorical Cross-Entropy Loss, Sparse Categorical Cross-
Entropy Loss, Hinge Loss (SVM Loss), Triplet Loss

Unit III: Optimizer function 3 Lecture Hours

Stochastic gradient descent, Mini-Batch Gradient Descent, Momentum, Adaptive
gradient algorithm (Adagrad), Adam (Adaptive Moment Estimation), RMSprop (Root
Mean Square Propagation), Adadelta

Unit IV: Data Preprocessing 3 Lecture Hours

Data Cleaning: handling Missing Data, Handling Outlier, Data Transformation: Feature
Scaling, Feature Encoding, Feature Engineering, Data Reduction: Dimensionality
reduction technique, feature selection, Data Splitting: Cross validation techniques,
Handling imbalanced data: Oversampling techniques, under sampling techniques.

Unit V: Regression 9 Lecture Hours

Introduction to Regression, Regression examples, Regression models, Steps in
regression analysis, Linear regression, Simple linear regression, Mathematical proof
of linear regression, Least squares estimation, Least squares regression-Line of best
fit, Illustration, Direct regression method, Maximum likelihood estimation, Coefficient
of determination (R-squared), Checking model adequacy, Over-fitting, Detecting over-
fit models: Cross validation, Logistic regression, Mathematical proof of logistic
regression, multiple linear regression, Multiple linear regression model building,
Mathematical proof of Multiple linear regression model, Interpretation of multiple linear
regression coefficients-Partial regression coefficients, Standardized regression
coefficients, Missing data, Validation of multiple regression model, regularization,
ridge and lasso regularization.

Unit VI: Classification 12 Lecture Hours

Introduction, ML classifier, Classification and general approach, Classification
algorithms, Instance based learning, K-Nearest neighbour, Decision trees, Attribute
selection measure: Information gain, ID3 algorithm, Converting a tree to rules,
Bayesian algorithms, Ensemble, Ensemble of classifiers, Bagging, Boosting, Random
forests, Neural networks, Activation functions, Feedforward neural network, Multi-layer
perceptron, Back propagation algorithm, Recurrent or feedback architecture,
Perceptron rule, Multilayer networks and back propagation algorithm, Support vector
machine, Classification model evaluation and selection, ROC curves, AUC curves.

Unit VII: Clustering Techniques 12 Lecture Hours

Introduction to Clustering, Clustering algorithms, Statistics associated with cluster
analysis, General applications of clustering, Clustering as a pre-processing tool,
Similarity and dissimilarity between objects, Type of data in clustering analysis, Binary
variables, Nominal variables, Ordinal variables, Cluster centroid and distances,
Hierarchical clustering, Hierarchical Agglomerative Clustering (HAC), Hierarchical
Agglomerative Clustering: Linkage method, Hierarchical Agglomerative Clustering:
Variance and Centroid method, Cluster distance measures, agglomerative clustering,
Distance between two clusters, Hierarchical clustering: Time and Space requirements,
K - means clustering, The K-medoids clustering method, CLARA (Clustering Large
Applications), Density based clustering methods, DBSCAN.

Total lecture Hours 45

References*
Textbooks 1. Andreas C. Müller and Sarah Guido,
"Introduction to Machine Learning
with Python", Shroff/O'Reilly, 2016.
2. Christopher M. Bishop , "Pattern
Recognition and Machine Learning",
Springer, 2016.
Reference books 1. Kevin P. Murphy, "Machine Learning:
A Probabilistic Perspective", MIT
Press, 2012.
2. Sebastian Raschka and Vahid
Mirjalili, "Python Machine Learning",
2nd Edition, Packt Publishing, 2017.

Web Resources
Journals
MOOCs, online courses

Modes of Evaluation: Quiz/Assignment/ presentation/ extempore/ Written

Examination

Examination Scheme
Components IA MID SEM End Sem Total
Weightage (%) 50 20 30 100
 Applied Machine Learning Lab

List of Experiments
Below is a list of small machine learning-based projects suitable for this lab work.
Through these projects students are expected to implement the concepts of data
preprocessing and machine learning algorithms. These projects cover various
machine learning techniques and can serve as valuable learning experiences:

Experiment 1 Predicting Housing Prices: Develop a regression model to predict

house prices based on features like location, size, and amenities.
Experiment 2 Iris Flower Classification: Use the Iris dataset to build a
classification model that predicts the species of iris flowers.
Experiment 3 Handwritten Digit Recognition: Implement a digit recognition
system using the MNIST dataset and a neural network.
Experiment 4 Breast Cancer Diagnosis: Develop a breast cancer classification
model using medical imaging data (e.g., mammograms).
Experiment 5 Sentiment Analysis: Create a sentiment analysis tool that
classifies text reviews as positive or negative using natural language processing
(NLP) techniques.
Experiment 6 Spam Email Detection: Build a spam email filter using text
classification algorithms.
Experiment 7 Predicting Stock Prices: Develop a time series prediction model
to forecast stock prices.
Experiment 8 Credit Risk Assessment: Build a credit scoring model to assess
the creditworthiness of applicants using historical financial data.
Experiment 9 Recommendation System: Create a movie or book
recommendation system based on user behavior data (collaborative or content-
based).
Experiment 10 Anomaly Detection: Implement an anomaly detection system for
detecting outliers in data (e.g., fraud detection).
Experiment 11 Customer Churn Prediction: Develop a model to predict customer
churn in a subscription-based business.
Experiment 12 Fake News Detection: Create a model to classify news articles as
real or fake based on their content.
Experiment 13 Disease Diagnosis from Medical Images: Use medical imaging
data (e.g., X-rays) to diagnose diseases or conditions.
Experiment 14 Traffic Sign Recognition: Build a model that can recognize and
classify traffic signs in images or video streams.

Total Lab hours 30

References*
Textbooks 1. Andreas C. Müller and Sarah Guido,
"Introduction to Machine Learning
with Python", Shroff/O'Reilly, 2016.
2. Christopher M. Bishop , "Pattern
Recognition and Machine Learning",
Springer, 2016.

Reference books 1. Kevin P. Murphy, "Machine

Learning: A Probabilistic
Perspective", MIT Press, 2012.
2. Sebastian Raschka and Vahid
Mirjalili, "Python Machine Learning",
2nd Edition, Packt Publishing, 2017.

Web Resources
Journals
MOOCs, online courses

Modes of Evaluation: Quiz/Assignment/ presentation/ extempore/ Written

Examination

Examination Scheme: Continuous Assessment

Components Quiz & Viva Performance & Lab Report

Weightage (%) 50 % 50 %