Data Science and Machine Learning

Course Code Course Title Type Credit Prerequisite

Python for Data Statistics,
DEML0201 Science and Machine Practice 4 Problem Solving
Learning Using Python
DEML0203 Supervised and
Unsupervised Practice 4 Nil
Machine Learning
DEML0204 Computational Nil
Practice 4
Intelligence
DEML0202 ML Techniques in Nil
Practice 4
Image Processing
DEML0205 Advanced Neural Nil
Network and Deep Practice 4
Learning Techniques
DEET0800 Internship Practice 4
DEET0300 Project Project 6
Total 30

Page |1
 Python for Data Science and Machine Learning

Course Pre-requisite
Code Course Title T-P-PJ
Type
DEML0201 Python for Practice 0-4-0 Statistics, Problem
Data Science Solving Using Python
and Machine
Learning

Objective:
• Learn core Python scripting elements such as variables and flow control structures
• Understand the concepts of file I/O
• Plot data using appropriate Python visualization libraries such as Numpy, Matplotlib
and SciPy etc
Learning Outcome:
• Able to use object oriented concept to solve problems
• To quickly and easily draw plot or visualize the information through visualization
technique
• Write an error free program of minimum 200 lines of code

Evaluation Systems:

Internal Examination Component % of Marks Method of Assessment

Internal Practice 50 (40+10) Lab work + Learning Record

External Examination External Practice 50 Lab work

Total 100

Course Outline:

Module I (12 Hrs)

Familiarization of Python: Overview of Numbers, Array, Tuples, List, Dictionary, Date
and Time. Decision Making Statement, Iteration Statement, Break and Continue statement.
Function, File, Module.

Module II (15 Hrs)

NumPy: Arrays, Array indexing, Datatypes, Array math operation, Sequences, Repetitions
and Random number, Broadcasting, Polynomial, unique items and the counts.

Module III (15 Hrs)

SciPy: Numpy VS SciPy, File input/output, Special Function, Linear Algebra, Operation,
Interpolation, Optimization and fit, Statistics and random numbers, Numerical Integration,
Fast Fourier transforms, Image manipulation.

Module IV (20 Hrs)

Plotting using Matplotlib : Basic plot(), Histogram, Bar Plot, Box Plots, Area Plot, Scatter
Plot, Pie Chart.

Page |2
Data Visualization: Visualization Libraries, Data frame: Data types, Attributes, methods
(mean(), median(), std(), var(), cor(), min()/ max(), describe()), groupby method, Selecting
Column, Filtering, Selecting row and column, Missing values, Data read from excel, CSV
and txt file.

Module V (12 Hrs)

Pandas: Use of Pandas. Pandas Installation, Data frame, Series, Create Data frame, Range
Data, Inspecting data, Slice data, Drop a column, Concatenation, Use pandas to handle Excel
Files.
Module VI (10 Hrs)
OpenCV: What is Computer Vision? Install OpenCV, Reads an image, Basics of OpenCV.
Environment, Storing images, Reading images, Writing images.
Module VII (12 Hrs)
Use SciKit-Learn for Machine Learning: Scikit-learn, Download and Install scikit-learn,
Machine learning with scikit-learn, Import the data, Create the train/test set.
References

Web Resource:

1. https://docs.python.org/3/tutorial/
2. https://www.analyticsvidhya.com/blog/2015/04/comprehensive-guide-data-exploration-
sas-using-python-numpy-scipy-matplotlib-pandas/
3. https://www.dataquest.io/blog/sci-kit-learn-tutorial/

Page |3
 Supervised and Unsupervised Machine Learning

Course
Code Course Title T-P-PJ Pre-requisite
Type
Supervised and
Unsupervised
DEML0203 Machine Practice 0-4-0 Nil
Learning

Objective:

• To provide a strong foundation of fundamental concepts in statistics and techniques of

artificial Intelligence and machine learning.
• To develop skills of using python program for solving practical problems in machine
learning.

Course Outcome:
• To able to design and implement machine learning solutions to classification, regression,
and clustering problems and be able to evaluate and interpret the results of the
algorithms.

Evaluation Systems:
Internal Examination Component % of Marks Method of Assessment
Internal Practice 50 (40+10) Lab work + Learning Record

External Examination External Practice 50 Lab work

Total 100

Course Outline:
Pre-requisite:
Measures of Central tendency: Mean, Median, Mode Geometric Mean and Harmonic Mean,
Use of Measures of Central tendency, Measures of Dispersion.
Correlation and Regression bivariate and multivariate
Module I (08 Hrs)
Machine Learning: Introduction, what is Machine Learning? , Types and Applications of
ML, AI vs. ML, Essential Math for ML and AI.
Meaning and Definition of Statistics-Role of Statistical Techniques in Decision Making.
Measures of Central Tendency: Mean, Median & Mode - Measurement of Dispersion: Mean
Deviation & Standard Deviation.

Module II (07 Hrs)

Regression Analysis: Types of regression, Linear Regression, Multiple linear regressions,
Logistic regression. Binary Classification Using Neural Networks, Multi-Class Classification.

Module III (12 Hrs)

Supervised Learning: Supervised Learning, Linear Methods for Classification, Linear
Methods for Regression, Support Vector Machines, Basis Expansions, Model Selection
Procedures.

Page |4
.

Module IV (12 hrs)

Unsupervised Learning: Brief to Unsupervised Learning, Association Rules, Cluster
Analysis, Reinforcement Learning, K-Means Clustering Project, Hierarchical Clustering.

Module V (15 Hrs)

Neural Networks: Introduction to Neural Networks, Classification of Neural network,
Neural Network Elements and Application, Single neuron as a linear classifier, commonly
used activation functions, Architecture of neural networks (Feed-forward networks, Feedback
networks, Network layers). , Representational power, Setting number of layers and their sizes.

Module VI (12 Hrs)

Neural Net Notes 2:Setting up the data and the model: Data Preprocessing, Weight
Initialization, Batch Normalization,. Loss function, Train/val accuracy, Weights:Updates
ratio, Activation/Gradient distributions per layer.

Module VII (12 Hrs)

Case Study: Projects

References

Web Resource:

1. https://www.coursera.org/learn/machine-learning-with-python
2. https://towardsdatascience.com/beginners-guide-to-machine-learning-with-python-
b9ff35bc9c51
3. https://www.python-course.eu/machine_learning.php

Page |5
 Computational Intelligence

Code Course Title Course Type T-P-PJ Pre-requisite

Computational
DEML0204 Intelligence Practice 0-4-0
Nil

Objective:

• Fundamentals of key intelligent systems technologies including knowledge based

systems, neural networks, fuzzy systems, and evolutionary computation, and
• Practice in integration of intelligent systems technologies for engineering applications.

Course Outcome:

• Gain a working knowledge of knowledge-based systems neural networks, fuzzy systems,

and evolutionary computation.
• Apply intelligent systems technologies in a variety of engineering applications.

Evaluation Systems:
Internal Examination Component % of Marks Method of Assessment
Internal Practice 50 (40+10) Lab work + Learning Record

External Examination External Practice 50 Lab work

Total 100

Course Outline:

Module I (15Hrs)
Introduction To Neural Network & Back propagation: Introduction , Simple expressions,
interpreting the gradient ,Compound expressions, chain rule, BackpropagationIntuitive
understanding of Backpropagation , Modularity: Sigmoid example , Backprop in practice:
Staged computation , Patterns in backward flow , Gradients for vectorized operations ,
Summary , Linear Backprop Example.

Module II (10Hrs)
Introduction To Fuzzy Logic Principles
Basic concepts of fuzzy set theory – operations of fuzzy sets – properties of fuzzy sets – Crisp
relations – Fuzzy relational equations – operations on fuzzy relations – fuzzy systems – propositional
logic – Inference – Predicate Logic – Inference in predicate logic – fuzzy logic principles – fuzzy
quantifiers – fuzzy inference – fuzzy rule based systems – fuzzification and defuzzification – types.

Module III (10Hrs)

Advanced Fuzzy Logic Applications
Fuzzy logic controllers – principles – review of control systems theory – various industrial

Page |6
applications of FLC adaptive fuzzy systems – fuzzy decision making – Multiobjective decision
making – fuzzy classification – means clustering – fuzzy pattern recognition – image processing
applications – systactic recognition – fuzzy optimization.

Module IV (10Hrs)
Evolutionary Algorithms
Introduction to Optimization, the Basic (Binary) Genetic Algorithm, Advanced Genetic Algorithms,
Optimization using Genetic Algorithms, Advanced Applications of GA, Types of GA, Mathematical
Foundations of GA, Genetic Programming, Particle Swarm & Ant Colony Algorithms

Module V (15 Hrs)

Hybrid Soft Computing Techniques, Hybrid system, neural Networks, fuzzy logic and Genetic
algorithms hybrids. Genetic Algorithm based Back propagation Networks: GA based weight
determination applications: Fuzzy logic controlled genetic Algorithms soft computing tools,
Applications.

References

Web Resource:

1. http://cs231n.stanford.edu/slides/2017/cs231n_2017_lecture6.pdf
2. https://www.coursehero.com/file/27112423/Neural-Nets-1pdf/
3. https://www.coursehero.com/file/27990916/Lecture-6-Notes-Gradient-Calculationpdf/
4. https://medium.com/@sidereal/cnns-architectures-lenet-alexnet-vgg-googlenet-resnet-
and-more-666091488df5

Page |7
 ML Techniques in Image Processing

Course Pre-requisite
Code Course Title T-P-PJ
Type
DEML0202
ML Techniques
in Image Practice 0-4-0
Nil
Processing

Objective:
• To study and learn different techniques for image enhancement, video and image
recovery.
• To study different techniques for image and video segmentation.
• To study different techniques for image and video compression and object recognition.

Learning Outcome:
• Able to Implement different Real life Projects in the filled of medical and Agricultural
image processing.

Evaluation Systems:

Internal Examination Component % of Marks Method of Assessment

Internal Practice 50 (40+10) Lab work + Learning Record

External Examination External Practice 50 Lab work

Total 100

Course Outline:

Module I (12 Hrs)

Digital Image and Video Fundamentals: Digital image and video fundamentals and
formats, image decimation/interpolation, video sampling and interpolation, Basic image
processing operations, Image Transforms using OpenCV.

Module II (15 Hrs)

Image and Video Enhancement and Restoration: Histogram, Point processing, filtering,
image restoration, algorithms for 2-D motion estimation, change detection, motion-
compensated filtering, frame rate conversion, deinterlacing, video resolution enhancement,
Image and Video restoration (recovery).

Module III (15 Hrs)

Feature Extraction: Image representations (continuous and discrete), Edge detection, Edge
linking, corner detection, texture, binary shape analysis, boundary pattern analysis, circle and
ellipse detection, Introduction to SIFT and SURF,ORB.

Page |8
Module IV (20 Hrs)
Image and Video Segmentation: Discontinuity based segmentation- Line detection, edge
detection, thresholding, Region based segmentation, Scene Change Detection, Spatiotemporal
Change Detection, Motion Segmentation, Simultaneous Motion Estimation and Segmentation
Semantic Video Object Segmentation.

Module V (12 Hrs)

Object recognition: Image Feature representation and description-boundary representation,
boundary descriptors, regional descriptors, feature selection techniques, introduction to
classification, supervised and unsupervised learning, Template matching, Bayes classifier

Module VI (10 Hrs)

Case Study 1-
Implementation of Face Reorganization System.

Module VII (12 Hrs)

Case Study 2-
Implementation of Diabetic Retinopathy prediction algorithm.

References

Web Resource:
1. https://dzone.com/articles/opencv-python-tutorial-computer-vision-using-openc
2. https://docs.opencv.org/3.0beta/doc/py_tutorials/py_feature2d/py_table_of_contents_fea
ture2d/py_table_of_contents_feature2d.html

Page |9
 Advanced Neural Network and Deep Learning Techniques

Course Prerequisite
Code Course Title T-P-PJ
Type
Advanced Neural
Network and Deep
DEML0205 Practice 0-4-0
Learning Nil
Techniques
Objective:

• To Design and train a CNN, RNN language models

• To Compare quality of CNN vs RNN

Course Outcome:

• Understand the general principles of deep learning methods

• Be able to apply deep learning principles using tools such as TensorFlow, Keras

Evaluation Systems:
Internal Examination Component % of Marks Method of Assessment
Internal Practice 50 (40+10) Lab work + Learning Record

External Examination External Practice 50 Lab work

Total 100

Course Outline:

Module I(15 Hrs)

Recurrent Neural Networks:Introduction to Recurrent Neural Networks,Theano Basics,
Theano Neural Network in Code, Tensorflow Basics, Tensorflow Neural Network in Code.
The Simple Recurrent Unit : Architecture of a Recurrent Unit,Unfolding a Recurrent Network,
Backpropagation Through Time (BPTT) ,The Parity Problem , The Parity Problem using a
Feedforward ANN, The Parity Problem using a Recurrent Neural Network,On Adding
Complexity

Module II(12Hrs)
Advanced RNN Units:Rated RNN Unit,RRNN in Code - Revisiting Poetry Generation,
Gated Recurrent Unit (GRU), GRU in Code, Long Short-Term Memory (LSTM).

Module III(12Hrs)
Convolutional Neural Networks: Architecture Overview
ConvNet Layers: Convolutional Layer, Pooling Layer, Normalization Layer, Fully-
Connected Layer, Converting Fully-Connected Layers to Convolutional Layers.
ConvNet Architectures: Layer Patterns, Layer Sizing Patterns,Case Studies (LeNet /
AlexNet / ZFNet / GoogLeNet / VGGNet), Computational Considerations.
Regularization (L2/L1/Maxnorm/Dropout)

P a g e | 10
Module VI(12Hrs)
Keras Overview: Backend in Keras, Theano in KerasTensorflow, and CNTK Backend
Comparing the Backends, KerasvsTensorflow, Advantages of Keras, Installing Keras
Direct install or Virtual Environment, Amazon Web Service (AWS), Amazon SageMaker
Training using the Fashion-MNIST dataset and Keras.

Module V(12 Hrs)

BatchTraining,TensorFlow, Basics Review:Introducing Tensors,Plane
Vectors,Tensors,Installing TensorFlow,Loading And Exploring The Data,Feature
Extraction,Deep Learning With TensorFlow,Running The Neural Network.
Module VI (10Hrs)
CNN Architectures:AlexNet, VGG, GoogLeNet, ResNet

Module VI(12Hrs)
Neural Nets: Gradient checks, Sanity checks, Babysitting the learning process:
Visualization.
Parameter updates: First-order (SGD), momentum, Nesterov momentum, Annealing the
learning rate, Second-order methods, Per-parameter adaptive learning rates (Adagrad,
RMSProp).Hyperparameter Optimization, Evaluation: Model Ensembles.

Module VII(12Hrs)
Using Keras to solve 5 assignments, Using Tensor Flow To solve 5 assignments

Books:
Text Book:
1. Learn Computer Vision Using OpenCVWith Deep Learning CNNs and RNNs,
SunilaGollapudi and V laxmikanth, APRESS,2019 Edition
2. LearnimgTensorFlow, Tom Hope, YehezkelS.Resheff, and Itay Lieder ,O’REILLY, 2017
1ST edition
Web References:
3. https://www.tensorflow.org/tutorials/
4. https://www.pyimagesearch.com/2018/09/10/keras-tutorial-how-to-get-started-with-
keras-deep-learning-and-python/

Project Areas:
• Financial marketing Analysis
• Plant disease detection
• Rainfall prediction
• Credit card fraud detection
• Real-Time Face Detection and Recognition for Security Purposes
• Data Mining Methods for Traffic Accident Severity Prediction
• Medical decision making diagnosis
• Real-time vehicle detection and tracking
• Analyze weather conditions and its prediction using deep learning techniques.
• Sales Prediction and Forecasting
• Recommender Systems for movieslist
• Social Media Sentiment Analysis using Twitter Dataset
• Boston Housing Price Prediction
• Human Activity Recognition using Smartphone Dataset

P a g e | 11
• Predict Stock Prices
• Uber Helpful Customer Support
• ML in Image processing
• Facial expression detector
• Currency Identification System
• Intelligent Traffic Light Control
• Diabetic Retinopathy prediction

P a g e | 12