Module Title: Data Analytics and Mining

Module Code: CSMDM21

Type of Assignment: coursework

Individual / Group Assignment: Individual

Weighting of the Assignment: 100%

Page limit/Word count: Maximum 16 pages

.     Excluding the front page of information and references.

.     Including figures, diagrams, graphs, and tables.

.    Times New Roman, 12pt., 1.15 line spacing.

.    Table of content, abstract/introduction are not required.

.    The report should be clearly structured with a separate section (with appropriate subsection) for each task and a final conclusion.

Expected hours spent for this assignment: 40 hours

Items to be submitted on-line through Blackboard Learn:

1. report.pdf,

2. workflow_group.knar - exported KNIME workflow group containing data files and workflow(s).

Work to be submitted on-line via Blackboard Learn by: 2023 November 24 (Friday) 12:00 noon

Work will be marked and returned by: 2023 December 15 (Friday) 12:00 noon

Note

By submitting this work you are certifying that you have read the assessment guidelines which are displayed at the top of the Assessment Folder on the Blackboard course for this module, and that you have conformed to the

associated policies and practises, including those on

.    Submitting your own work, not that of other people or systems, and the associated penalties for Academic Misconduct

.    Submitting by the specified deadline, and the penalties associated with late submission (if allowed) .    The exceptional circumstances system (for applying for extensions)

.    The use of a green sticker for students with relevant needs

ASSIGNMENT DESCRIPTION

Major Coursework (100% of module assessment)

This assignment should be carried out using KNIME - the data analytics and mining platform presented in the module.

In this assignment, you are required to build and test classification models for a customer dataset to predict the target customer segmentation.

.     Data: customer.csv (available on Blackboard)

.     Description: customer_description.txt (available on Blackboard)

Task 1: Data Understanding and Preprocessing (30%)

Construct a KNIME workflow to understand the data characteristics and quality, report and discuss your findings. Based on the data understanding, identify and discuss the required data preprocessing steps, and perform them in the KNIME workflow.

Task 2: Classification (30%)

Next, add to or construct another KNIME workflow to build at least two classification models using the dataset by experimenting with at least two different algorithms and/or their hyperparameters. You can use any classification algorithms. In the report, describe the adopted algorithms, also discuss and justify your selection of algorithms and parameters. You may use experiments to support your discussions and justifications.

Task 3: Model Evaluation (30%)

Finally, add to the KNIME workflow to evaluate the trained models using appropriate performance measures and evaluation methods. In the report, describe the adopted performance measures and evaluation methods, discuss and justify your selection of performance measures and evaluation methods, present and analyse the results, also discuss the result reliability.

Report Quality (10%)

The  report  should  be  clearly  structured  with  a  separate  section/subsection  for  each  task/subtask  with  a  final conclusion  and  references.   For  each  task,  solutions  should  be  described  with justifications/discussions.   KNIME workflow  images  should  be  presented  in  the  report,  including  relevant  node  configurations.  Results  should  be presented in the report with analysis and discussions. Sections, figures and tables must be numbered. References should    follow    a    suitable     academic    format    (https://www.reading.ac.uk/library/finding-info/guides/lib-citing- references.aspx).