Javelin1991 / CZ4041_Machine_Learning

Year-4 Module taken in NTU that discusses about various machine learning algorithms and their strengths and weaknesses

CZ4041 : Machine Learning

Learning Outcome

I have learnt how various machine learning methods or algorithms work as well as their strengths and weaknesses. The coursework involves selecting a given topic/task on Kaggle and then implementing the best possible machine learning model for the selected task.

Coursework Project - Leaf Classification

• Implemented 8 different machine learning models to perform leaf classification - Naive Bayes Classifier, Support Vector Machine, K-Nearest Neighbours, Logistic Regression,Linear Discriminant Analysis,Decision Tree,Random Forest and Multilayer Perceptron
• Analyzed the performance of the above machine learnings algorithms for leaf classification
• MLP produces the best prediction results (i.e. the highest accuracy and the lowest log loss value)
• Submitted the prediction results to Kaggle
• Placed among top 15% on the public leaderboard according to the final evaluation scores on Kaggle.

For the project demonstration video :
https://www.youtube.com/watch?v=pQM_NAij_qY&feature=youtu.be

Disclaimer : The python code used to implement the coursework project is no longer maintained. There may be errors or bugs that did not exist at the time of creation.

Follow the instructions below to run the source code.

Leaf Classification

Need to install using pip the following libraries:

1. numpy
2. pandas
3. seaborn
4. matplotlib
5. opencv-python
6. scikit-learn

Running the program

1. Type the command "python main.py" to run the program.
2. Once the program runs, follow the instructions to run a specific Machine Learning Method
3. To replicate the results submitted, select MLP to reproduce the submission file.

Knowledge Accquired includes:

1. Overview of machine learning and its applications

2. Decision Theory and Bayes Models

3. Classifier Evaluation

4. Classification: Decision trees, Artificial neural networks, linear and kernelized Support vector machines, K-nearest neighbour classifiers

5. Linear regression and its kernelized extension

6. Ensemble Learning

7. Clustering

8. Dimension Reduction

9. Density Estimation