Course Brief

  • To show how Business Intelligence (BI) tools, methodologies and techniques can be used to provide intelligent and efficient support for decision making.
  • To provide hands on experience in the tools and processes involved and a critical appreciation of the business and project management context and role of BI.
  • To critically study the concepts, principles and theory of Data Mining.
  • To provide practical experience of DM on a variety of datasets, in particular in the domain of BI.
  • To provide students with the ability to use, compare and select appropriate DM tools and algorithms/models to interpret and critically evaluate the results.
  • This course is offered in 60 hours

On completing this course successfully the students are expected to be able to:

  • Understand the theoretical underpinnings of BI and DM methodologies, architectures, techniques and algorithms.
  • Conduct an audit and analysis of the BI requirements of an organisation and contribute to the planning of a BI project as part of a Knowledge Management.
  • Critically appraise the Business process change requirements, and analyse/design, implement and evaluate the key elements of BI projects, including HCI, Business Reports and other aspects of building a successful BI system.
  • Critically evaluate and select appropriate DM facilities, algorithms/models and apply them and interpret and report the output.
  • Critically appraise the design and implementation of a DM application/technology using test/sample but realistic data sets and modern tools.
  • Integrate intelligent and DM elements into a BI systems development project.

Introduction to BI and DM: The overall picture of BI and DM; Integrating DM components into BI systems.
BI analysis and Audit methods and tools; Architectures for BI; Data Warehousing.
Management Information Systems; Methodologies for BI projects; CRISP.
HCI, explanation and visualization.
Approaches for evaluating the value of BI systems to the business.
Data preparation/Data cleaning – Dealing with missing values, outliers and erroneous data.
Machine Learning/Training and statistical fitting: supervised and unsupervised learning.
Models over cases – Unsupervised (Clustering, ARM): distance and similarity measures, K-means, Kohonen Neural Net (Self-Organizing-Maps (SOMs)), Case-based reasoning approach; Supervised (prediction models): Classification (Decision trees, Multiple Discriminant Analysis), Linear regression models, Logistic regression for risk modelling, Multi-Layer-Perceptron, Regression trees.
Classification accuracy and the confusion matrix.
Models over attributes – Measures of association, the odds-ratio, the correlation matrix.
Mining the cube for associations, the priori algorithm.
Efficiency, granularity and scalability of BI and DM systems.

Concepts will be introduced during lectures, and developed during tutorial and lab sessions. Extensive use will be made of case-study materials when students will be expected to analyse problem situations, present their findings and suggest courses of action.

Student time will be:
Lecture 50%
Tutorial 25%
Laboratory 25%

Larger Coursework – 100%
5000 words.
Answer all questions.
Pass mark – 50%