Dr Ata Kaban -- Room 32. Final Year & MSc project topics

 

News:

  Recent student projects published at the UKCI 2013 conference:

Sachin Mylavarapu and Ata Kaban. Random projections versus random feature selection for classification of high dimensional data. Proceedings of the UK Workshop on Computational Intelligence (UKCI 2013), September 9-11, University of Surrey, Guilford, UK - accepted.

Zhenghui Ma and Ata Kaban. K-Nearest-Neighbours with a novel similarity measure for intrusion detection. Proceedings of the UK Workshop on Computational Intelligence (UKCI 2013), September 9-11, University of Surrey, Guilford, UK - accepted.



Topic areas for projects:

* Data mining, visualisation

-          data projections (dimensionality reduction)

-          web log mining, text mining

-          visualising time series

-          outlier detection (e.g. to find peculiar quasars in astronomy)

-          independent components analysis

 

* Music projects

-          retrieval by tapped rhythm

-          composition (e.g. using evolutionary techniques)

-          style recognition (e.g. using machine learning)

-          harmonisation (e.g. using a Hidden Markov Model)

 

* Image analysis

-          face recognition (e.g. using PCA+LDA)

-          image segmentation (e.g. using clustering)

-          image super-resolution (e.g. using optimisation)

-          image-text translation (image annotation)

 

* Bioinformatics

-          gene network recovery (e.g. using screening+regression)

-          gene or protein array classification

 

* Educational software

-          interactive teaching for Machine Learning

 

* Research oriented projects

- talk to me if interested (or find out what I do)

 

* Students' own idea is most welcome

 

Expectations:

-          enthusiasm for your project

-          good programming skills

-          willingness to work hard

-          willingness to learn some new techniques [specially if you aim for a high mark!]

 

 

A useful pointer to concepts in Machine Learning

http://alumni.media.mit.edu/~tpminka/statlearn/glossary/glossary.html

 

 

Examples of past projects:

 

Machine learning for intrusion detection

Detecting faces in a still image

Automatic harmonisation of Bach chorales

Missing value prediction for galaxy spectra

Decision trees: An online interactive teaching tool

Image segmentation using entropy minimisation

Melody retrieval by tapped rhythm