Machine Learning
|
Supervisor |
|
|
Suitable for |
MSc in Computer Science
|
Abstract
Dr. Vasile Palade (vasile.palade@cs.ox.ac.uk) will supervise projects in machine learning, ranging from more theoretical approaches to practical applications of machine learning. A preference will be given for projects on neural networks, support vector machines, evolutionary computation and fuzzy techniques applications to problems from the bioinfomatics area, web usage mining, financial modelling, but other options will be considered too (e.g. image processing).
For example, one possible project may be to investigate the performance of several variants of particle swarm optimization algorithms on classical benchmark optimization problems, such as the CEC 2005 benchmark suite, which can then be applied to some real-world optimization problems from the above areas. Another project may be to study the problem of class imbalance learning, and develop some classifiers or ensemble of classifiers using some of the machine learning methods listed above, in the presence of highly imbalanced datasets. Popular imbalanced data sets, such as those from the UCI machine learning repository, but not only, could be used.
These projects would be more appropriate for MSc and 4th year students, but 3rd year students could choose such a project if they have built an appropriate background through the courses they had previously attended. Ideally, students should have attended the Machine Learning course, or at least the Intelligent Systems or the Information Retrieval course. Please feel free to contact Dr. V. Palade and discuss about possible projects in these areas.