CMP7239 - Presentation: Practical Machine Learning: Applied Machine Learning Assessment Brief 2024-25 S2 | BCU

Assessment Summary

This is an individual assessment that requires students to implement practical machine learning tools using a software platform and evaluate the effectiveness of the tools to achieve the task’s objectives.

The students need to synthesize and communicate the findings using a recorded video presentation along with the result’s report document.

Assessment Details:

Title: Presentation – Practical Machine Learning
Theme: Applications related to Computer Networks, or Cybersecurity (depending on the enrolled course)
Style: Recorded video presentation and presentation slides.

Rationale:

Guided by technological advances in computing, communication and electronics, Artificial Intelligence (AI) has gained momentum in progressing the development of intelligent systems over the past decade. Machine learning is at the core of the recent wide-spread AI deployment and adoption, enabling machines to be trained to perform automated intelligent tasks. Whilst computer science has been the primary area of machine learning development and applications, recent trends have shown machine learning applications in wider domains of networking, cybersecurity, and smart systems.

Hands-on knowledge and skills in applying machine learning concepts and building learning models are key to demonstrating the effective intelligent data-driven solution to domainspecific information processing challenges. This assessment aims to support the development of relevant knowledge and skills, focusing on the practical machine learning task using a software tool.

Description:

This is an individual assessment that requires you to submit a recorded video presentation and report document to explain findings from practical machine learning experiments and empirical evaluation.

• You can use the datasets from assessment 1 or any other dataset that aligns with your course (networks, cybersecurity, or smart systems). Sources that might help you find suitable datasets:

o https://www.kaggle.com/datasets

o https://ieee-dataport.org/

o https://dl.acm.org/artifacts/dataset

o https://www.journals.elsevier.com/data-in-brief

• You have to develop machine learning model(s) in WEKA that combines feature engineering and other supervised and unsupervised techniques (Preferably 2 or more algorithms for results and evaluation). The models have to be validated using standard performance metrics such as accuracy, precious, recall and F1 score 
• The number of machine learning models to be developed depends on how the performance benchmark is obtained.
  
  o Option 1: Benchmark performance metrics that are obtained from a published research article. For this option, only one model is required, and a comparison is to be made between your model and the performance of any published article (preferably the selected paper from Assessment 1).
  o Option 2: Two models are developed where one of them is the benchmark.

Learning Outcomes

LO4 - Synthesize and communicate findings from an empirical investigation to diverse backgrounds