Machine Learning

Paper Code: 
25DBCA501C
Credits: 
03
Periods/week: 
03
Max. Marks: 
100.00
Objective: 

Course Objectives:

The objective of course is to:

1.   Introduce students to basic applications, concepts, and techniques of machine               learning.

2.   Develop skills in students to implement machine learning algorithms on real world        problems and evaluate their performance.

 

Course Outcomes: 

Course

Learning Outcome

(at courselevel)

Learning and

teaching

strategies

Assessment

Strategies

Course

Code

Course

Title

25DBCA 501C

 

Machine Learning (Theory)

CO313. Formulate a problem for data analytics and analyses data and select suitable machine learning technique for designing a model.

CO314. Develop a machine learning model for unlabeled data using association rule.

CO315. Develop a machine learning model for unlabeled data using clustering technique.

CO316 Analyse and implement different classification techniques.

CO317. Analyse and implement different prediction techniques. Compare their performance.

CO318 Contribute effectively  in course- specific interaction.

Approach in teaching:

Interactive Lectures, Discussion, Reading assignments, Demonstration.

 

Learning activities for the students: Self learning assignments, Effective questions, Seminar presentation.

Class test, Semester end examinations, Quiz, Assignments, Presentation.

 

9.00
Unit I: 

Introduction to Data Mining and machine learning: Basic Data Mining Tasks, Data Mining versus Knowledge Discovery in Databases, Applications of Machine Learning, Machine Learning vs AI, Types of Machine Learning, Metrics, Accuracy Measures: Precision, recall, F-measure, confusion matrix, cross-validation.

 

9.00
Unit II: 

Understand the Problem by Understanding the Data: unbalanced data, Unsupervised Learning: Association rules, Apriori algorithm, FP tree algorithm, and their implementation in Python, Market Basket Analysis and Association Analysis.

 

9.00
Unit III: 

Clustering: k-means and implementation of k-means using python, Concept of other clustering algorithms: Expectation Maximization (M) algorithm, Hierarchical clustering, and DBSCAN.

 

9.00
Unit IV: 

Classification & Prediction: model Construction, performance, attribute selection Issues: under, Over-fitting, cross validation, tree pruning methods, missing values, Information Gain, Gain Ratio, Gini Index, continuous classes. Classification and Regression Trees (CART) and C 5.0. Implementation of decision tree in python

 

9.00

Classification  &  Prediction:  Linear  Regression,  Multiple  Linear  Regression, Logistic Regression,  Naïve  Bayes  and  Support  Vector Machines(SVM), Implementation  of  Linear Regression, Logistic Regression, Naïve Bayes and SVM in python.

ESSENTIAL READINGS: 

Essential Readings:

1.  Jiawei Han & Micheline Kamber, “Data Mining: Concepts & Techniques”, Morgan            Kaufmann Publishers, Third Edition.

2.  Sebastian Raschka & Vahid Mirjalili,” Python Machine Learning”, Second                       Edition,Packt>.

3.  McKinney, Python for Data Analysis. O’ Reilly Publication,2017.

 

REFERENCES: 

SUGGESTED READINGS:

1.  Miller, Curtis. Hands-On Data Analysis with NumPy and Pandas: Implement               Python Packages from Data Manipulation to Processing. United Kingdom: Packt          Publishing, 2018. (Latest editions of the above books are to be referred)

e RESOURCES:

1.  https://www.jigsawacademy.com/blogs/business-analytics/

2.  NOC: Python for Data Science, IIT Madras                                                             ,https://nptel.ac.in/courses/106106212

3.  Python, w3scool, https://www.w3schools.com/

4.  Jupiter :www.jupiter.com

5.  Googlecolab: www.googlecolab.com

JOURNALS:

1.  Journal of Machine Learning Research (JMLR),ACM, https://dl.acm.org/journal/jmlr

2.  International Journal of Machine Learning and Cybernetics, springer :

https://www.springer.com/journal/13042

 

Academic Year: