B.Engg(ME), M.Sc.(Computing)Contact details for Mr Rashid Mahmood
Thesis
Interactive Optimisation based Protein Structure PredictionDescription
Interactive optimisation allows human decision makers, typically problem domain experts, to contribute to the optimisation process. They can perform candidate solution evaluation in cases where no mathematical model exists. They can also assume a guiding role in an optimisation process where the constraints and objectives either are mathematically intractable or are tractable but the complexity of the problem domain renders autonomous approaches inefficient.
In a problem domain such as protein structure prediction it is a simple observation that the search space is infeasible large for many search algorithms. Further, the knowledge of the processes involved is currently incomplete and is based on rudimentary models. Therefore, interactive guidance may hold the potential of significant improvement of solution quality and reduced computation times.
The automation process will involve employing machine learning approaches to derive domain specific rules from the decision makers' selections. The automation process might also lead to insights into the problem itself.
Supervisors
Professor Abdul Sattar
Dr Tamjidul Hoque
Research expertise
- Computational Intelligence
- Search Algorithms
- Data Mining
- RDBMS
- Computational Biology
- Bioinformatics
- Computer Programming
Publications
- M. A. Rashid, Md Tamjidul Hoque and Abdul Sattar, “Association Rules Mining Based Diseases Correlation Prediction”, Medinfo2010- 13th World Congress on Medical and Health Informatics, 2010 (Cape Town, South Africa)