Timeline for Identify outliers for annotation in text data
Current License: CC BY-SA 4.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 26, 2020 at 21:14 | comment | added | Erwan | @MykolaZotko ok, probably for this purpose just counting the matches is sufficient. But in general this approach would favor long documents over short ones, that's why I think normalizing is better (and it's not really more complex). | |
| Dec 26, 2020 at 20:59 | vote | accept | Mykola Zotko | ||
| Dec 26, 2020 at 20:59 | comment | added | Mykola Zotko | I checked the source code. For each word in a document, you need to count all matches in training (labeled) data, sum them up, and divide by the number of words in the document. This should be a simple approach as the author says. | |
| Dec 26, 2020 at 18:39 | history | answered | Erwan | CC BY-SA 4.0 |