 | Politics Project‑class | ||||||
| |||||||
Hello everyone! I'm currently working on a project with Aaron Halfaker where we are studying how to do automated classification of article importance. In case you're wondering who I am and why I'm the one working on this, I've added a short explanation of that at the end of this post.
One of our project's goals is to make article importance predictions for articles within a specific WikiProject, and we have been making good progress towards that goal. We are now reaching out to a handful of WikiProjects to learn more about how useful and accurate our models are. These projects were chosen because almost all of them mention what the "average Wikipedia reader" is likely to look up as a criterion when determining article importance, a concept that corresponds to our usage of article views as input in our model, and because they have a large number of articles rated as "unknown importance", something we can build tools to assist with. WikiProject Politics is one of these projects, which is why I'm posting here.
During the past week I have been working on building a model that predicts the importance ratings of articles within the scope of WikiProject Politics. The model uses data such as number of article views, number of inlinks from other articles (and what proportion of those come from within the WikiProject), as well as information from the clickstream dataset. I have created a page listing almost 100 articles where our model believes the importance rating of a given article should change, you can find that list here: User:SuggestBot/WikiProject Politics#Candidates for re-rating
We'd love to know whether our predictions are accurate or useful. For instance, you might find that a prediction is not spot on, but enough to warrant changing the rating of a given article (e.g. we might predict Top-importance for an article that should be High-importance, but if it is currently rated Low-importance both cases warrant an update of that rating). If our predictions are wrong, please let us know about that too! There might be aspects of importance that our model does not catch and we want to understand what those are so we can see if there are ways to fix them.
While working on building this model, I also identified many pages within the scope of WikiProject Politics that appear to need an update of their importance rating. These pages are also listed on the page linked above. For example there are many redirects and disambiguation pages that do not have an NA ("not an article") rating, and I've updated the rating of some of them. I have also identified some articles that do not appear to have a matching Wikidata item, or where the Wikidata item is labelled a "Wikimedia disambiguation page".
As I mentioned in the introduction, in case you’re wondering who I am and why I'm the one working on this: as part of my PhD I studied using machine learning to predict article quality, and I helped develop parts of ORES to enable it to make article quality predictions. Besides doing research I also run SuggestBot, a bot which recommends you articles to edit, currently available in seven languages. Thanks for your time, and I’m looking forward to your thoughts and questions! Regards, Nettrom (talk) 16:25, 11 May 2017 (UTC)[reply]
Julius Caesar is currently rated high, but the assessment for "high" is, "has not achieved international notability, or is only notable within a particular continent". I'm pretty sure that he has achieved international notability and is notable maybe in every continent. I mean he was from Rome, Europe and we are talking about him here in America, where he is probably extremely notable also. Thinker78 (talk) 05:50, 22 March 2018 (UTC)[reply]
The example for an FA article was demoted last month. A new article is needed. --Aknell4 (talk • contribs) 18:02, 8 May 2021 (UTC)[reply]
I wonder how relevant is parties-and-elections.eu as source of used frequently on Wikipedia a sit seems to be a website of single individal enthusiast? *(claims to be lawyer from Frankfurt) --Zblace (talk) 16:31, 26 January 2022 (UTC)[reply]
