JuLiA is an automated comment moderation system. Uses natural language processing to parse for the semantics of abusiveness. **ENGLISH ONLY**
JuLiA stands for Just a Linguistic Algorithm. This plugin checks your comments against the JuLiA web service, which supplies an abusiveness score for each comment. Based on this score, comments are auto-published, auto-deleted, or marked for review. Over time, scores are aggregated by commenter in order to generate abusiveness rankings and other statistics.
When the plugin is activated, it will add an "Abusive" category to the top of your comments management screen. When JuLiA parses a comment and decides that it is abusive, it will be removed from your "Pending" category and placed into the "Abusive" category. In addition, when JuLiA decides that a comment is definitely clean then it will simply be auto-published and save you time by skipping the "Pending" category entirely.
JuLiA makes these decisions based not on keywords, but on the overall semantics present in a single comment. For now, JuLiA works on English comments only.
A new top-level "JuLiA" menu will be added to the sidebar of your admin screen, where you can change plugin settings, view your most abusive commenters, and set your list of Flagged and Trusted commenters. Many of these settings are self-explanatory, but the most important are the Auto-Publish and Quarantine thresholds.
Whenever JuLiA parses a comment, she assigns it a score ranging from 100% Abusive to 100% Clean. By default, JuLiA will Auto-Publish a comment if it scores higher than 50% Clean, and will Quarantine a comment if it is more than 80% Abusive. All other comments are sent to the "Pending" category with a recommendation attached. These thresholds can be changed in the "Settings" sub-menu.
Over time, JuLiA scores are aggregated and grouped by commenter to produce a very accurate measure of your most abusive commenters. By default JuLiA will list the most abusive commenters for the past 7 days, but this timeframe can be changed in the Settings menu. You have the option to flag commenters who appear in the abusiveness report, and you can manage flagged and trusted commenters in a different sub-menu.
When a commenter is Flagged as abusive, their comments will never be Auto-Published regardless of how they score. When a commenter is marked as Trusted, their comments will always be Auto-Published regardless of how they score. By default, your own username is automatically added to the Trusted list. You can add or remove users from this list in the appropriate sub-menus.