Group project

2 - Project output (other than dissertation) - 50

Individual coursework

1 - Written assignment (including essay) - 50

To view information for a specific offering, select a value from the drop down above.

A selection of topics will be covered and some examples are as follows:  
<ul><li>Bias-variance trade-off;  </li><li>Loss function and knowledge about standard numerical algorithms to solve convex and non-convex optimisation problems; </li><li>Evaluation metrics to validate the supervised learning algorithms; </li><li>Cross-validation to tune the hyperparameters; </li><li>Classification algorithms: </li></ul>
<ol><li>Discriminative approaches, for example, k nearest neighbours, support vector machine, tree-based models; </li><li>Generative approaches, for example, Naïve Bayes, linear/quadratic discriminant analysis. </li></ol>

Feedback is provided in a number of ways throughout your module. This includes feedback you receive during teaching from both peers and lecturers and that provided for formative assessment. Following an assessment, you will be given your marks and feedback in line with the assessment regulations and policy. More information on the timing and type of feedback that will be provided for each assessment will be available from the module leader. <a href='https://www.citystgeorges.ac.uk/about/governance/policies/student-policies-and-regulations'>https://www.citystgeorges.ac.uk/about/governance/policies/student-policies-and-regulations</a>

This module will support the development of all the graduate attributes. Your module leader will provide you with information about how your learning, teaching and assessment activities support your development of these attributes. The graduate attributes are: - Always learning - Engaged in the world - Technical and digital - Connected Professional - Creating Impact

In-Person - 

Self-directed - 

The Elements of Statistical Learning.

 An Introduction to Statistical Learning: with Applications in Python.

redictive Analytics and Data Mining: Concepts and Practice with RapidMiner. 

 Fundamentals of Machine Learning for Predictive Data Analytics: Algorithms, Worked Examples, and Case Studies

Supervised Machine Learning

Container 1

BA2103 - Supervised Machine Learning

BA2103-PRD1-PER-ALV1

A variety of learning and teaching methods will be used in this module. 
Lectures are used to introduce context, concepts and techniques illustrated with practical and current examples. You will also have the opportunity to participate in class discussions and work through examples and exercises with the support of the lecturer.  It is strongly recommended that you attend all lectures. 
In addition, there will be regularly scheduled computer labs. The latter are intended to demonstrate the use of technology for the practical implementation of the material covered during the lectures using observational data. The module is highly participative. You are also expected to devote an equivalent amount of learning time through individual and group study on a weekly basis. 
Key learning and teaching resources will be put on the module website on Moodle.  
Self-directed learning hours are for independent study and account for the minimum amount of time you should spend studying independently.  For example, carrying out subject research, practicing and reflecting on techniques learned in class, reading widely around the subject to develop a deeper understanding, carrying out activities using the VLE, attending induction and career development workshops, working in groups to complete assignments and presentations, and completing individual assignments and other homework.

Find suitable hyperparameters for supervised learning algorithms to achieve good test performance; 

Skills

Apply appropriate methodologies to understand the relationship between the response and explanatory variables; 

Knowledge

Illustrate the fundamental concepts of supervised learning and the differences between supervised and unsupervised machine learning; 

Explain and discuss the different supervised learning approaches  to deal with data driven real-world problems; 

Recognise ethical considerations in the application of supervised learning, such as bias, privacy and fairness.  

Values and attitudes

Apply creativity and adaptability to novel scenarios where supervised learning can provide insights. 

Identify different ways to evaluate the performances of the classification algorithms; 

Demonstrate awareness of the potential consequences of misusing supervised learning algorithms. 

Demonstrate a critical view of supervised learning algorithms.

Analyse real-world data using supervised learning; Implement supervised learning algorithms using a programming language; 

Apply appropriate methodologies for out-of-sample prediction; 

Select and explain the choice of specific supervised techniques for different data analysis tasks. 

Apply appropriate algorithms to solve small, medium and large sized optimisation problems  

Explain the main idea and the methodologies of supervised learning algorithms;  

Explain and critically analyse the output of the supervised learning algorithms; 

Examine the quality of supervised learning results using appropriate evaluation metrics. 