How to select an Evolutionary algorithm (like GA, PSO, ACo etc.) among the lot of existing ones for feature optimization? I mean what are the parameters which we need to see when we select any such algorithm for any pattern classification problem?
I am attaching you works devote to an analysis and synthesis of engineering systems methodology and to the integration of variables method, that could help you to carry out a correct formulation of your problem and to define the appropriate solution method
Primary reasons for feature selection are:
1. Dimensionality reduction -->speeding up the system
2. reducing model complexity
Where speed is not a major concern, the model complexity can be controlled by
controlling the regularization parameter.
Hi Ram,
There are numerous of algorithms in R for feature selection and you can use GUI to use them. Also, deep learning with python contains some neural network example. In addition, you can ECJ or jmetal packages for evolutionary algorithms. Also, you can check https://searchcode.com/ and http://www.artificial-immune-systems.org/algorithms.shtml for immune system. Furthermore, nature-inspired algorithms in https://books.google.co.uk/books?id=iVB_ETlh4ogC&printsec=frontcover&dq=meta+algorithms+oxford&hl=en&sa=X&redir_esc=y#v=onepage&q&f=false and computational intelligence in https://www.crcpress.com/Computational-Intelligence-Paradigms-Theory--Applications-using-MATLAB/Sumathi-Paneerselvam/p/book/9781439809020.
Best wishes,
Osman
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology