Dr. Firat Tekiner - Home Page

PhD Projects Available!

PhD projects are available for exceptionally good candidates in Computer Science/Engineering for EU/UK Nationals in the areas of High Performance Computing, Data Mining and AI applications on MANET routing.

For more information contact Dr. Firat Tekiner

Project Ideas

Project suggestions for UCLAN Computing students. Use the projects below as a guidance to prepare your prroject proposals after discussing with me.

Data Mining and Programming Projects

Project 1 - Programming GPGPU for general computing

• Aim: exploit the performance of GPGPUs for general computing purposes.
• Application Area: Data Mining
• Goal: Port a data mining application to GPGPU and compare performance against traditional processors Tool: NVIDIA CUDA Environment
• Required Skills: C/C++ Programming skills
• Performance Criteria: Time to process given dataset

Project 2 - Programming PS3 for general computing

• Aim: exploit the performance of IBM Playstation3 for general computing purposes.
• Application Area: Data Mining
• Goal: Port a data mining application to GPGPU and compare performance against traditional processors
• Tool: IBM Play station Development environment
• Required Skills: C/C++ Programming skills
• Performance Criteria: Time to process given dataset

Project 3 - Data Mining using Weka

• Aim: Compare performance of a number of Data Mining Algorithms.
• Goal: Investigate the performance and capabilities of different algorithms and applications?
• Simulation Tool: WEKA (freely available)
• Applications: freely available datasets on the net
• Issues: Data Cleaning, Data processing, visualisation
• Performance Criteria: Accuracy, Speed, etc

Project Suggestion 4 - Parallel Data Mining

• Aim: Investigate Parallel computing approaches
• Goal: Take a serial data mining application and produce a parallel version
• Simulation Environment: UCLAN HPC Cluster
• Applications: freely available datasets on the net
• Issues: MPI programming
• Performance Criteria: Scalability, Speed of processing

Project Suggestion 5 - Parallel Data Mining - Survey

• Aim: Investigate Parallel computing approaches
• Goal: Produce a very detailed comparison of the existing approaches for producing parallel data mining applications
• Issues: No simulation or practical implementation is required. However, each approach will be investigated in very detailed manner. It is expected that classification of approaches produced as well as comparison based on the criteria defined within the thesis.