|
75 | 75 | of higher density.</td> |
76 | 76 | </tr> |
77 | 77 | <tr> |
78 | | - <td>eGPmix</td> |
| 78 | + <td>GPmixture</td> |
79 | 79 | <td>Functional</td> |
80 | 80 | <td>Here</td> |
81 | | - <td> </td> |
82 | | - <td> ... </td> |
| 81 | + <td><a href="https://github.com/mingz628/GPmixture" target="_blank">GitHub</a></td> |
| 82 | + <td> GPmixture is for learning mixtures of Gaussian processes. The idea |
| 83 | + is to project the functional data into a few orthonormal functions, |
| 84 | + perform cluster analysis of the projection coefficients for each |
| 85 | + orthonormal fuction, and aggregate individual clusterings into a |
| 86 | + concensus clustering.</td> |
83 | 87 | </tr> |
84 | 88 | <tr> |
85 | 89 | <td>FAE</td> |
|
93 | 97 | <td>Multivariate</td> |
94 | 98 | <td>Here</td> |
95 | 99 | <td><a href="https://pypi.org/project/CPFcluster/" target="_blank">PyPI</a></td> |
96 | | - <td> ... </td> |
| 100 | + <td> Component-wise Peak-Finding (CPF) is an improvement over DCF: (1) |
| 101 | + the assignment methodology is improved by applying the density peaks |
| 102 | + methodology within level sets of the estimated density; (2) the |
| 103 | + algorithm is not affected by spurious maxima of the density and hence is |
| 104 | + competent at automatically deciding the correct number of clusters.</td> |
97 | 105 | </tr> |
98 | 106 |
|
99 | 107 | </table> |
|
0 commit comments