FAQ
How frequently is the data updated and/or generated?
The algorithm itself runs non-stop, crawling the web and enriching it's knowledge representation. Updates to this interface are essentially only limited by the underlying data, which must exhibit a significant p-value before it is selected for posting. Intuitively, a slow news day should lead to a slow BS day, but in reality this is generally not true.
Does the algorithm analyze post comments?
Our first post to this interface was a comment which scored off the charts in BS significance, but we soon noticed that by allowing the analysis of comments we introduced so much variability into the system that it would eventually become polluted and have a difficult time distinguishing even the most blatant falsehoods. Moreover, since our focus is on the way memes spread from side-stream sources, we have decided to disable comment processing.
Are post titles generated by the algorithm?
Unfortunately no, once the BS detector identifies the top hits for a specific time period, we enter them into this web interface and try to choose titles that are appropriate or humorous. The titles are not meant to be part of the analysis, but are there simply for entertainment.
What is a Subgraf?
Often when a major new event occurs, many related nodes will appear simultaneously, making identification of spurious data difficult. In these situations, it can still be informative to simply look at the most representative nodes. We present the results of such an analysis under the Subgraf section (from the news-speak term "nut graf" and sub-graphs in graph theory).
What's up with the algorithm name?
Who says scientists can't have a sense of humor?