UFCFMJ 15 M
UFCFMJ-15-M Machine Learning and Predictive Analytics
Overview
This module introduces the theory and practice behind a range of supervised and unsupervised machine learning methods, together with software implementations and best practice in evaluation and performance measurement.
From a foundation in principles of machine learning, this module aims to equip students with a flexible toolkit of methods and techniques, culminating in a demonstration of their abilities to apply them appropriately and effectively to a problem of their own choice.
Objectives
- Synthesise evidence on the value of data as an asset for businesses to “mine” knowledge and “predict” trends
- Develop and evaluate predictive analytics approaches and techniques such as regression and random forest classifiers
- Apply problem solving skills necessary for identifying the organisational capacity needed to employ a predictive analytics solution
Curriculum
Indicative list of topics:
Introduction to predictive analytics: Defining predictive analytics - introduction
Business Relevance of PA - Business intelligence and applications: Relevance of pattern recognition, classification, optimisation
Predictive analytics and big data Case study: a business application using predictive analytics approaches
Predictive analytics in business - applications: Sources of data and value of knowledge
Identify a wide range of applications for predictive analytics: Marketing and recommender systems, fraud detection, business process analytics, credit risk modelling, web analytics and others Social media and human behaviour analytics
Case study: email targeting - which message will a customer answer? - (tutorial)
Analytics models and techniques: Introduction to analytics modelling
Types of analytics models: Predictive models Survival models Descriptive models
Define pattern recognition, inferring data and data visualisation Briefing learning and regression approaches Comparison of approaches - use and goals - (tutorial)
Introduction to machine learning: Introduction: Basic principles: Basic notions of learning Introduction to learning problems (classification, clustering and reinforcement) and literature Identifying different learning approaches - supervised, unsupervised and reinforcement
Case study on different types of learning - (tutorial)
Machine learning for predictive analytics (1): Review of types of problems
Machine Learning techniques: Decision tree learning Artificial neural networks Clustering Naive Bayes classifier k-nearest neighbours Genetic algorithms
Case study on problem - a “suitable” predictive modelling technique - (tutorial)
Regression techniques for predictive analytics: Review of types of problems (application) Linear regression models Survival or duration analysis (time to event analysis) Ensemble learning and random forest Case study on problem - a “suitable” predictive modelling technique - (tutorial)
Advanced topics and Software tools: Analytics in the context of big data Predictive analytics as art and science Software tools; the R project and Python
Trends and challenges in predictive analytics - where are we going?
Assessment
Report (100%)