Timeline for Why do I get weird results when using high perpexity in t-SNE?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 12, 2018 at 9:46 | vote | accept | Mathias Andersen | ||
| Mar 8, 2018 at 20:37 | history | edited | amoeba | CC BY-SA 3.0 |
added 6 characters in body; edited title
|
| Mar 8, 2018 at 14:27 | answer | added | amoeba | timeline score: 3 | |
| Mar 8, 2018 at 14:10 | comment | added | amoeba | Then I am surprised that scikit-learn allows to run t-sne with this parameter at all. I might add this as an answer a bit later today. | |
| Mar 8, 2018 at 13:32 | history | edited | Mathias Andersen | CC BY-SA 3.0 |
added 201 characters in body
|
| Mar 8, 2018 at 13:01 | history | edited | Mathias Andersen | CC BY-SA 3.0 |
edited title
|
| Mar 8, 2018 at 12:57 | comment | added | Mathias Andersen | I see. n is 300 in these examples as well. | |
| Mar 8, 2018 at 11:27 | comment | added | amoeba | What is the sample size here? Linked scikit code uses n=300, and you can't use perplexity larger than n: it's not possible to achieve it. Some tsne implementations will not allow it at all. | |
| Mar 8, 2018 at 10:57 | history | asked | Mathias Andersen | CC BY-SA 3.0 |