| Category |
Assignment |
Subject |
Computer Science |
| University |
University of London (UOL) |
Module Title |
OUBS035312 Machine Learning |
OUBS035312 Assignment 1
Identify and describe 5 applications of machine learning in the education sector. For each application, provide a detailed explanation and include examples of how it benefits students, teachers, or educational institutions. Include relevant references.
OUBS035312 Assignment 2
Your task is to select a dataset from Kaggle that involves a multi-class classification problem. You will apply three different supervised learning algorithms to solve the problem. You are required to document and explain every step of the machine learning process, from data preprocessing to model evaluation. This assignment aims to assess your understanding of the machine learning workflow and your ability to implement and optimize machine learning models for multi-class classification.
Assignment Requirements:
1. Dataset Selection:
- Choose a multi-class classification dataset from Kaggle. Ensure the dataset has more than two classes in the target variable.
- Provide a brief description of the dataset, including its features, target variable, and any relevant context.
2. Preprocessing:
- Handling Missing Values: Describe how you handled any missing data in the dataset and justify your approach.
- Normalization: Explain how you normalized the data, and why it was necessary for your selected algorithms.
- Feature Selection: Identify and select the most relevant features for your models. Explain the technique you used for feature selection.
- Handling Imbalanced Classes: If your dataset has imbalanced classes, explain the technique you used to address this issue.
3. Model Implementation:
- Apply three different supervised learning algorithms suitable for multi-class classification.
- Split the data into training and testing sets. Justify your choice of split ratio.
- Train each model on the training data and test it on the testing data.
4. Model Optimization:
- Discuss the hyperparameter tuning process for each model and the methods you used.
- Compare the performance of the models before and after optimization, focusing on how well they handle the multi-class problem.
5. Model Evaluation:
- Explain the evaluation metrics used for multi-class classification (e.g., Accuracy, Precision, Recall, F1-Score and Confusion Matrix).
- Provide a comparative analysis of the models based on these metrics, discussing which model performed best and why.
6. Documentation:
- Submit a detailed report explaining every step of the machine learning process. Your report should include:
- Dataset description
- Data preprocessing steps
- Explanation of each algorithm and its suitability for multi-class classification
- Model training and testing processes
- Handling of imbalanced classes (if applicable)
- Hyperparameter tuning process
- Evaluation and comparison of model performance
- Include clear and well-commented code snippets within the report.
7. Assumptions:
- Clearly state any assumptions made
8. References:
- Include relevant references.
9. Source Code:
- Submit the complete source code as a separate file/notebook. Ensure that your code is well-organised and documented with comments.
OUBS035312 Submission Guidelines:
- The source code should be in Interactive Python Notebook(.ipynb)
- All files should be compressed into a single ZIP folder for submission on Blackboard.
