ICT5357 Problem-Solving and Decision-Making with Machine Learning

Assessment 1: Supervised Machine Learning Project

Assessment Overview

Weighting: 20%
Due Date: Week 4
Assessment Type: Individual assessment
Word Count / Length: 800 (650 words for report + equiv. 150 model)

Unit Learning Outcomes
In this assessment, you will be tested on whether you have successfully met the following Unit Learning Outcomes (ULOs):
• ULO1: Apply strategic problem-solving to achieve goals, complete tasks, make decisions with data
• ULO2: Research supervised learning and unsupervised learning techniques and apply appropriate techniques to solve problems that involve classifying or clustering data
• ULO4: Plan, design and create machine learning models to solve a variety of problems, using automated machine learning (AutoML).

Submission Type and Required Format
The type of assessment you will be completing is an Unsupervised Machine Learning Project.
It should include the following:
• Title and Introduction – Clearly state the dataset chosen and a brief overview of the problem.
• Problem Definition – Describe what problem could be addressed with the dataset and its significance.
• Machine Learning Approach – Specify whether it is a classification or regression task, and what the model will do.
• Algorithm Justification – Explain suitable/unsuitable algorithms with reasons.
• Data Preprocessing – Describe cleaning, transformation, and feature engineering steps.
• Model Evaluation – Present performance metrics and interpretation.
• References – Follow APA 7th edition.
• Other Requirements: Include trained and tested model(s) file.

Assessment Details

Assessment Purpose
The purpose of this assessment is:
• Enhance skills in identifying practical machine learning problems from raw datasets.
• Apply AI algorithms for classification or regression task.
• Develop proficiency in data preprocessing, model training, and evaluation.
• Strengthen analytical and problem-solving skills in AI project design and deployment.
• Prepare students for real-world AI/ML application development.

Assessment Instructions
In this assessment, you will need to complete the following steps:

1. Choose one dataset from the options provided: https://apic.instructure.com/courses/2056/assignments/33024
2. Define a problem that could be solved with the dataset.
3. Explain how the dataset relates to the problem and whether the task is classification or regression.
4. Select a suitable ML algorithm, justifying why it is appropriate (and noting unsuitable options).
5. Describe preprocessing steps you would take before model training.
6. Train at least one model using the prepared dataset.
7. Evaluate model performance with relevant metrics.
8. Submit:
   • A written report.
   • The project file(s) used to train and test the model(s).
   Make sure your work aligns with the Unit Learning Outcomes (ULO1, ULO2, and ULO4), demonstrates critical thinking, and follows the format guidelines provided.

Artificial Intelligence (AI) Use
Permitted AI Use:
• Brainstorming ML project ideas and refining the problem statement.
• Structuring the report or creating outlines.
• Grammar checking and improving clarity.

Not Permitted:
Using AI to generate the entire assessment without personal input.
Copying AI-generated content without validation from reliable sources.

Responsibility: Students must ensure accuracy and integrity of all content.
Declaration: Any AI use must be acknowledged in the AI Declaration on the cover sheet.

Other Important Information

Assessment Resources
• ECA Library and Learning Support.
• Dataset Pack provided in the unit’s Canvas page.
• Unit Readings (ICT5357) and lecture notes.
• APA 7th Referencing Guide from ECA Library & Learning Support.
• An example ML project exercise will be conducted in class.

Submission Requirements
On completion of your assessment:
• Follow the Assessment Presentation Guidelines as outlined on the ECA Library and Learning Support website.
• Submit your document as a PDF/MS Word file, trained and tested model(s) file, and project source file via the assessment submission link in Canvas.
• Include a completed Assessment Cover Sheet, with the required AI Declaration, and attach it to your submission.
• Save drafts of your files in case you need to revise or resubmit.

Assessment Support
For academic support or feedback on a draft of your assessment, please email [email protected]
For assistance with finding resources, such as books and journal articles, please email [email protected]
For information and guides on tackling assessments and developing your academic skills, please visit in the ECA Library and Learning Support website: https://eca.libguides.com/
For queries about this specific assessment task, please contact the Unit Coordinator.

Assessment 1 Rubric

Criteria (weighted as indicated below)

High Distinction (HD) 85-100

Distinction (D) 75-84

Credit (C) 65-74

Pass (P) 50-64

Fail (F) 0-49

Criteria 1: Problem Solving (20 marks)

Able to identify and connect problems to machine learning models and training data and provide clear explanations for these connections.

Able to identify and connect problems to machine learning models and training data.

Able to identify and connect many problems to machine learning models and training data.

Able to identify and connect some problems to machine learning models and training data.

Unable to identify and connect problems to machine learning models and training data.

Criteria 2: Data (20 marks)

Explanation is exceptional, offering detailed insights into features, target variable, and dataset's role in the task. Modifications or pre-processing are exceptional, showing deep understanding of their impact on the task.

Explanation is thorough and demonstrates a strong grasp of dataset characteristics and relevance. Modifications or pre-processing are well-reasoned and effectively contribute to model development.

Explanation provides a clear overview of the dataset, including features, target variable, and relevance to the task. Modifications or pre-processing are adequately explained and justified.

Dataset is described with basic details but lacks clarity or depth. Basic modifications or pre-processing steps are mentioned, but their rationale is unclear.

Explanation of the dataset is missing or contains significant inaccuracies. No modifications or pre-processing are performed, or explanations are unclear.

Criteria 3: Machine Learning (20 marks)

Algorithm selection and hyperparameter tuning are exceptional, showing deep understanding of model choices.

Algorithms are well-suited to the task, and hyperparameters are carefully tuned with clear reasoning.

Appropriate algorithms are selected, relevance to the task is justified with reasonable hyperparameters.

Basic algorithms are chosen, but their suitability and hyperparameters are inadequately explained.

Inappropriate algorithms or hyperparameters are chosen, or explanations are missing.

Criteria 4: Models (20 marks)

Model performance is explained clearly; interpretation shows deep understanding.

Model outcomes are explained with insight; their implications are well interpreted.

Provides some meaningful insights, and results are described clearly.

Basic model results are mentioned but lack context or interpretation.

No model results provided, or results do not address the problem.

Criteria 5: Clarity and Presentation (20 marks)

Answers are clear and organised, easy to understand, and engaging or insightful.

Answers are clear and organised.

Answers are mostly clear and organised.

Answers are somewhat clear but lack organisation.

Answers are unclear, disorganised, and difficult to understand.