File:ORES damaging calibration errors by newcomers and anons.png

Original file ‎(1,150 × 703 pixels, file size: 52 KB, MIME type: image/png)

This is a file from the Wikimedia Commons. The description on its description page there is copied below.

Summary

Description	English: ORES is a machine learning system for classifying edits to Wikipedia according to whether they are damaging or goodfaith. This plot shows that, for the 25 Wikis on which ORES damaging models are deployed, and when the prediction threshold is chosen to maximize overall accuracy, that the model tends to underestimate the incidence of damaging edits. However, the models are most conservative for editors that are neither newcomers nor anonymous. Compared to other editors, the models score edits by newcomers and anons are more likely to be damaging.
Date	23 April 2019
Source	Own work
Author	Groceryheist

I, the copyright holder of this work, hereby publish it under the following license:

You are free:

Under the following conditions:

attribution – You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
share alike – If you remix, transform, or build upon the material, you must distribute your contributions under the same or compatible license as the original.

Click on a date/time to view the file as it appeared at that time.

	Date/Time	Dimensions	User	Comment
current	21:30, 3 May 2019	1,150 × 703 (52 KB)	Groceryheist	decrease resolution a bit for legibility.
	16:53, 2 May 2019	1,460 × 703 (54 KB)	Groceryheist	Change title.
	22:20, 25 April 2019	1,460 × 679 (46 KB)	Groceryheist	Change y axis label and dodge the pointranges.
	18:44, 24 April 2019	1,150 × 679 (45 KB)	Groceryheist	Show 95% confidence intervals around the points.
	05:42, 24 April 2019	1,150 × 679 (30 KB)	Groceryheist	User created page with UploadWizard