Timeline for Cubic Clustering Criterion using R update
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 13, 2017 at 12:44 | history | edited | CommunityBot |
replaced http://stats.stackexchange.com/ with https://stats.stackexchange.com/
|
|
| May 15, 2016 at 21:22 | vote | accept | Pierre L | ||
| May 13, 2016 at 12:26 | answer | added | Pierre L | timeline score: 4 | |
| May 13, 2016 at 11:17 | review | Close votes | |||
| May 13, 2016 at 15:05 | |||||
| May 12, 2016 at 22:09 | history | edited | Pierre L | CC BY-SA 3.0 |
added 372 characters in body
|
| May 12, 2016 at 21:21 | comment | added | Pierre L | Can you say that in English? : ) Is the kernel bandwidth and uniform sampling thresholds in the R output or input parameters? | |
| May 12, 2016 at 21:19 | comment | added | EngrStudent | The CCC is a comparison of how much better the clustering is than a uniform random. If you had a decent first estimator of kernel bandwidth, then you could use a kernel method to tell you "where your data is". Also called a "nonparametric density". You could restrict your uniform sampling to those regions if you use a threshold. (0.001 aka 0.1%?) It might require a stronger signal to show, but might also work against things like "the curse of dimensionality". | |
| May 12, 2016 at 19:56 | history | asked | Pierre L | CC BY-SA 3.0 |