Wikidata:Item quality/Modeling
This page is used for collecting the idea for engineering the features for item quality model.
References edit
An item can not be considered of high quality if many of its properties are not referenced.
Also, not all references are equal. A reference "imported from * Wikipedia" is not worth much (except for images). A reference from a reliable work is better than from a tabloid.
Image Quality edit
An item with a good image has better quality than an item with no image, but on the opposite a small blurry image does not add much to quality.
Image quality can be very subjective, but some metrics are pretty telling:
- Size
- Contrast
- Sharpness
This information could be obtained by using ImageMagick (command-line) and OpenCV (library).
See also edit
- Linked Data Quality of DBpedia, Freebase, OpenCyc, Wikidata, and YAGO. http://www.semantic-web-journal.net/system/files/swj1465.pdf
- Research:Wikidata_gap_analysis
- Utilizing the Wikidata System to Improve the Quality of Medical Content in Wikipedia in Diverse Languages: A Pilot Study - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4468594/