This is an implementation of a forward search inspired PSO algorihtm (FSIPSO) for feature selection (FS) for high-dimensional data. FSIPSO uses a forward search scheme to dynamically expand the search space of FS problems. The mutation operation is also used in FSIPSO to improve the search performance. For a detailed description of the method please refer to
Li, A.-D., Xue, B., & Zhang, M. (2021). A Forward Search Inspired Particle Swarm Optimization Algorithm for Feature Selection in Classification. IEEE Congress on Evolutionary Computation, CEC 2021, Kraków, Poland, June 28 - July 1, 2021, 786–793. https://doi.org/10.1109/CEC45853.2021.9504949 [BibTeX]
The source code is in the src folder. An illustration to run the FSIPSO algorithm is implemented in ./src/fs/RunFSIPSO.java. Please note that code requires the jar file (./lib/weka.jar) of Weka, Machine Learning Software in Java. Please make sure it is added in the libiray.