The performance of each particle is measured according to a pre-defined fitness function. best experience. Then the particle flies toward a new position according to equation (12). Table 1 PPORSFS&GAAR parameter settings additional parameter settings for their operation. These are given in Table 1 * Optimal solution ** Exclusive optimal solution Table 3 Reduct sizes found by Feature Selection algorithms Table 4 Classification results with different reducts 1: Number of rules; 2: Classification accuracy Table 5 Experimental Results by RSES Note: For the last three datasets (Soybean-small, Lung and DNA), RSES cannot find all reducts by the exhaustive algorithm as it requires too much memory. Instead, we use a fast genetic algorithm to find 10 reducts. Fig.2 (a) Generation 1 of PSO on dataset Exactly2 Fig.2 (b) Generation 2 of PSO on dataset Exactly2 Fig.2 (c) Generation 4 of PSO on dataset Exactly2 Fig.2 (d) Generation 12 of PSO on dataset Exactly2 Fig.2 (e) Generation 13 of PSO on dataset Exactly2 Fig.4 Evolution process of the global best on dataset Vote. (PSORSFS, GAAR, POSAR) given in Table 7 and Figure 4. The process of the particle swarms searching for optimal solutions for dataset Vote is Example 2 Appendix A (Experimental Examples) Mushroom is given in Table 8 and Figure 5. Soybean-small is given in Table 9 and Figure 6. given in Table 10 and Figure 7. The process of the particle swarms searching for optimal solutions for dataset Lung is Example 5 Table 11 PSO searching process on DNA
