Author Affiliations
School of Electronics and Information Engineering, Lanzhou Jiaotong University, Lanzhou, Gansu 730070, Chinashow less
Fig. 1. Convergence analysis of four clustering algorithms. (a) FCM; (b) K-means; (c) DPC; (d) JW-FCM
Fig. 2. Clustering results of four clustering algorithms on Spiral data set. (a) FCM; (b) K-means; (c) DPC; (d) JW-FCM
Fig. 3. Clustering results of four clustering algorithms on S1 data set. (a) FCM; (b) K-means; (c) DPC; (d) JW-FCM
Fig. 4. Clustering results of four clustering algorithms on ISquare2 data set. (a) FCM; (b) K-means; (c) DPC; (d) JW-FCM
Fig. 5. Comparison of ACC indicators
Fig. 6. Comparison of ARI indicators
Fig. 7. Robustness comparison on Wine data set
Fig. 8. Robustness comparison on Thyroid data set
Fig. 9. Robustness comparison on D31 data set
Fig. 10. Robustness comparison on S1 data set
Fig. 11. Robustness comparison on Isquare2 data set
Fig. 12. Robustness comparison on Spiral data set
Data set | Sample number | Attributes | Number of categories |
---|
Wine | 178 | 13 | 3 | Thyroid | 215 | 5 | 3 | D31S1Isquare2Spiral | 310050001741312 | 2222 | 311563 |
|
Table 1. Experimental data set properties
Data set | ACC | ARI | Entropy | |
---|
FCM | K-means | DPC | JW-FCM | FCM | K-means | DPC | JW-FCM | FCM | K-means | DPC | JW-FCM | |
---|
Wine | 0.408 | 0.456 | 0.441 | 0.898 | 0.007 | 0.006 | 0.012 | 0.854 | 0.150 | 0.250 | 0.160 | 0.100 | Thyroid | 0.615 | 0.653 | 0.524 | 0.750 | 0.168 | 0.196 | 0.075 | 0.698 | 0.600 | 0.780 | 0.470 | 0.420 | D31 | 0.875 | 0.358 | 0.846 | 0.993 | 0.654 | 0.324 | 0.756 | 0.882 | 0.320 | 0.320 | 0.230 | 0.180 | S1 | 0.657 | 0.685 | 0.876 | 1.000 | 0.598 | 0.698 | 0.897 | 1.000 | 0.760 | 0.890 | 0.760 | 0.630 | Isquare2 | 0.976 | 0.764 | 0.985 | 1.000 | 0.975 | 0.708 | 0.968 | 1.000 | 0.080 | 0.270 | 0.130 | 0.060 | Spiral | 0.356 | 0.405 | 0.529 | 1.000 | 0.003 | 0.011 | 0.276 | 1.000 | 0.960 | 0.960 | 0.660 | 0.460 | Mean | 0.648 | 0.554 | 0.700 | 0.940 | 0.401 | 0.324 | 0.497 | 0.906 | 0.478 | 0.578 | 0.402 | 0.308 |
|
Table 2. Performance comparison of four clustering algorithms on data set