Table 7 The comparison of execution time (in seconds) for AMA and ADMM-based convex clustering methods
Data set | AMA method | ADMM-based conv. clust. like method 4 workers | ADMM-based conv. clust like method 8 workers | ADMM-based conv. clust like method 10 workers | ADMM-based conv. clust. like method 20 workers | ADMM-based conv. clust. like method 25 workers |
|---|---|---|---|---|---|---|
\(1000 \times 3\) | 2.8 | 12.65 | 9.73 | 13.79 | 11.45 | 12.24 |
\(5000 \times 3\) | 17.55 | 38.38 | 19.95 | 16.45 | 19.35 | 14.15 |
\(10{,}000 \times 3\) | 45.46 | 96.0 | 51.17 | 40.84 | 37.3 | 22.01 |
\(5000 \times 5\) | 22.74 | 39.29 | 19.65 | 16.87 | 16.15 | 14.37 |
\(10{,}000 \times 10\) | 100.2 | 136.76 | 72.63 | 59.33 | 50.23 | 30.29 |
\(200{,}000 \times 3\) | N.A. | – | – | – | – | 564.1 |