ECS5388

Pattern recognition

Offered subject to sufficient enrolment

R A Jarvis

6 points · 39 hours of lectures and practical work · Irregular availability · Clayton

Objectives The student is expected to acquire knowledge about supervised and unsupervised pattern recognition methodology and to develop critical judgement concerning the strengths and weaknesses of the various approaches covered in the context of application requirements.

Synopsis Supervised pattern recognition including binary feature vector analysis, Bayesian minimal risk formulations, near neighbour rules, sequential pattern recognition, linguistic and fuzzy logic systems, neural networks, evidential reasoning systems; unsupervised pattern recognition including minimal spanning trees, shared near neighbour rules and self organising neural networks.

Assessment Examinations: 50% · Practical work: 40% · Seminar: 10%

Back to the 1999 Engineering Handbook