This course covers the preliminaries of machine learning required for undergraduate students. The following content will be covered in the course:

Module 1:
- Introduction to Machine Learning: Evolution of Machine Learning, Paradigms for ML, Learning by Rote, Learning by Induction, Reinforcement Learning, Types of Data, Matching, Stages in Machine Learning, Data Acquisition, Feature Engineering, Data Representation, Model Selection, Model Learning, Model Evaluation, Model Prediction, Search and Learning, Data Sets. [PDF]
Module 2:
- Nearest Neighbor-Based Models: Introduction to Proximity Measures, Distance Measures, Non-Metric Similarity Functions, Proximity Between Binary Patterns, Different Classification Algorithms Based on the Distance Measures, K-Nearest Neighbor Classifier, Radius Distance Nearest Neighbor Algorithm, KNN Regression, Performance of Classifiers, Performance of Regression Algorithms. [PDF]
Module 3:
- Models Based on Decision Trees: Decision Trees for Classification, Impurity Measures, Properties, Regression Based on Decision Trees, Bias–Variance Trade-off, Random Forests for Classification and Regression.
- The Bayes Classifier: Introduction to the Bayes Classifier, Bayes’ Rule and Inference, The Bayes Classifier and its Optimality, Multi-Class Classification | Class Conditional Independence and Naive Bayes Classifier (NBC).
Module 4:
- Linear Discriminants for Machine Learning: Introduction to Linear Discriminants, Linear Discriminants for Classification, Perceptron Classifier, Perceptron Learning Algorithm, Support Vector Machines, Linearly Non-Separable Case, Non-linear SVM, Kernel Trick, Logistic Regression, Linear Regression, Multi-Layer Perceptrons (MLPs), Backpropagation for Training an MLP.
Module 5:
- Other Types of Learning: Ensemble Learning- Bagging, Boosting, Stacking and its impact on bias and variance, AdaBoost, Gradient Boosting Machines, XGBoost. Reinforcement Learning - Introduction, Q Learning.

References

Subramanian Chandramouli, Saikat Dutt, Amit Kumar Das, “Machine Learning”, Pearson Education India ,1st edition,2015.
Tom M. Mitchell, “Machine Learning’, MGH, 1997.

Programming Assignments

Programming Assignment 1: Distance Measures

Objective

Implement different distance measures between two vectors without using Python libraries.

Tasks

Implement Euclidean Distance
Implement Manhattan Distance
Implement Chebyshev Distance
Implement Minkowski Distance (p = 3)
Compare the distances for different pairs of vectors.

Sample Input

Python

Run

x = [2, 4, 6]

y = [1, 3, 5]

Expected Output

Euclidean Distance: 1.732

Manhattan Distance: 3

Chebyshev Distance: 1

Minkowski (p=3): 1.442

Programming Assignment 2: Non-Metric Similarity Functions

Objective

Compute similarity instead of distance.

Tasks

Implement Cosine Similarity
Implement Jaccard Similarity for sets
Implement Dice Similarity Coefficient
Compare the results on sample data.

Dataset

Python

Run

A = {1,2,3,4}

B = {3,4,5,6}

Programming Assignment 3: Proximity Between Binary Patterns

Objective

Measure similarity between binary vectors.

Tasks

Implement Simple Matching Coefficient (SMC)
Implement Jaccard Coefficient
Implement Hamming Distance
Compare all measures on multiple binary patterns.

Example

Python

Run

x = [1,0,1,1,0,1]

y = [1,1,1,0,0,1]

Programming Assignment 4: Distance-Based Classification Algorithms

Objective

Compare different nearest-neighbor classifiers.

Tasks

Use the Iris dataset.
Implement 1-NN classifier.
Implement 3-NN classifier.
Implement 5-NN classifier.
Compare classification accuracy.

Output

Generate a table showing k vs accuracy.

Programming Assignment 5: K-Nearest Neighbor Classifier from Scratch

Objective

Implement KNN without using Scikit-learn.

Tasks

Load a dataset (Iris or Wine dataset).
Split into train and test sets.
Compute Euclidean distance manually.
Find the k nearest neighbors.
Predict the majority class.
Calculate accuracy.

Deliverables

Source code
Confusion matrix
Accuracy score

Programming Assignment 6: Radius Distance Nearest Neighbor Algorithm

Objective

Classify samples using a fixed radius instead of a fixed number of neighbors.

Tasks

Implement Radius Nearest Neighbor.
Test with different radius values.
Analyze cases where no neighbor exists within the radius.
Compare results with standard KNN.

Suggested Radius Values

Python

Run

[0.5, 1.0, 1.5, 2.0]

Programming Assignment 7: KNN Regression

Objective

Use KNN for predicting continuous values.

Dataset

Use California Housing or Diabetes dataset.

Tasks

Train KNN Regressor with different k values.
Predict target values.
Plot actual vs predicted values.
Compute regression metrics.

Metrics

Mean Absolute Error (MAE)
Mean Squared Error (MSE)
Root Mean Squared Error (RMSE)
R² Score

Programming Assignment 8: Performance Evaluation of Classifiers

Objective

Evaluate KNN classification performance.

Tasks

Train KNN classifier.
Compute:
- Accuracy
- Precision
- Recall
- F1-Score
Generate a Confusion Matrix.
Plot the confusion matrix using Matplotlib.

Extension

Compare k = 1, 3, 5, 7, 9.

Programming Assignment 9: Performance Evaluation of Regression Algorithms

Objective

Evaluate KNN regression performance.

Tasks

Train KNN Regressor.
Compute MAE, MSE, RMSE, and R².
Plot error values for different k.
Identify the optimal k.

Expected Graph

Programming Assignment 10: Comparative Study of Distance Measures in KNN

Objective

Study how different distance measures affect KNN performance.

Dataset

Use Breast Cancer Wisconsin Dataset.

Tasks

Train KNN using:
- Euclidean distance
- Manhattan distance
- Minkowski distance (p=3)
- Chebyshev distance
Evaluate accuracy, precision, recall, and F1-score.
Present results in tabular form.
Write a short report (2–3 pages) discussing which distance measure performs best and why.

Submission Link: https://forms.gle/dbmJfSoF5bCZ7ygn6

Page updated

Google Sites

Report abuse