Last week, I submitted the interim report towards my dissertation project as part of my BSc Computer Science studies, which introduces and discusses many of the key concepts that will form a great deal of my individual project.
Following my growing interest in AI and machine learning, my dissertation will be on the 'Evolutionary Generation of Classifiers using Genetic Programming'. As is evident from the name, the research I will be conducting is on the application of genetic programming (GP) to solve classification problems, specifically two classification problem domains (both accessed via the UCI machine learning repository):
- Breast Cancer Wisconsin (Diagnostic) Data Set
- Climate Model Simulation Crashes Data Set
These datasets will be used to experiment and demonstrate the ability for the proposed GP system to solve such classification problems.
In terms of how the GP system will be implemented, the research will explore and conclude on a range of techniques and approaches that can be incorporated into the GP system, such as the fitness function and generational operators (selection, crossover and mutation). This proposed work will be implemented using Java, and will harness the use of a library to assist in the GP development called Jenetics.
As I've found in my interim report, this will be a demanding and challenging project with a great many components involved, but as I've also found, it will be a hugely enjoyable and rewarding piece of work to perform and complete. I'm excited to complete this work, and I look forward to distributing my dissertation paper once it's (eventually) finished.
