Gene Expression Programming - a new machine learning technique for supervised and unsupervised classification
Lead Research Organisation:
Brunel University London
Department Name: Sch of Engineering and Design
Abstract
Many scientific, engineering and business fields such as genetics, medicine, environment science and engineering, physics, astronomy, finance, and marketing are facing common challenges in dealing with complex data for extracting field-specific knowledge. Efficient data analysis techniques are needed in order to intelligently assist the user in extracting this knowledge. This project will address this need using the basic ideas of a recently developed computer algorithm, Gene Expression Programming, for the development of novel evolutionary algorithms techniques and novel supervised and unsupervised data classification algorithms. The project will develop and exploit novel homologous genetic operators, and mechanisms to control the redundant information in the solutions provided by the algorithm in order to increase its efficiency. These developments will be combined with state-of-the-art statistical methods such as boosting learning in order to create efficient data classification algorithms.The methods and algorithms developed in the project will be implemented in software applications made available as open-source in order to maximize the spectrum of the beneficiaries of the project outcomes.
Organisations
People |
ORCID iD |
Liliana Teodorescu (Principal Investigator) |
Description | The algorithm developed novel data analysis techniques based on a specialised version of an Evolutionary algorithm and developed theoretical studies of this algorithm |
Exploitation Route | The findings are implemented in software applications which are in the process of being released as open source. They can be available not only to the specialist in the area of evolutionary algorithms but also to a large range of practitioners from scientists and engineers to financial and health service specialists. |
Sectors | Aerospace Defence and Marine Agriculture Food and Drink Chemicals Digital/Communication/Information Technologies (including Software) Financial Services and Management Consultancy Healthcare Manufacturing including Industrial Biotechology Pharmaceuticals and Medical Biotechnology Retail Transport |
Description | The findings were used mainly by advancing the field of evolutionary algorithm and related data analysis techniques. Some of the data analysis techniques developed were applied to particle physics data analysis. |
First Year Of Impact | 2008 |
Sector | Digital/Communication/Information Technologies (including Software) |
Impact Types | Cultural |
Title | Boosted GEP |
Description | The data analysis techniques is based on a hybrid algorithm which combines the Gene Expression Programming and the AdaBoost algorithms for classification problems. |
Type Of Material | Data analysis technique |
Provided To Others? | No |
Impact | This techniques provides improved solutions to a classification problems with a reduced number of iterations of the searching process of the solution space. |
Title | Enhanced GEP |
Description | The data analysis techniques extended the capabilities of the Gene Expression Programming algorithm by using an alternative representation of the candidate solution and a truncated evolution process of the candidate solution for classification problems. |
Type Of Material | Data analysis technique |
Year Produced | 2008 |
Provided To Others? | Yes |
Impact | The data analysis technique was used on experimental data from particle physics experiments providing alternative methods for separating signal from background in such experiments. |
Title | BGEP |
Description | The software implements the hybrid algorithm based on Gene Expression Programming and AdaBoost algorithms developed in this project. |
Type Of Technology | Software |
Year Produced | 2012 |
Impact | The impact is on this project only at this time. The software is in the process of being released as open source. |
Title | GEP |
Description | The software is a new implementation of the Gene Expression Programming algorithm. |
Type Of Technology | Software |
Year Produced | 2010 |
Impact | The impact is on this project only at this time. The software is in the process of being released as open source. |