Preload
11-TheAlgorithm_Header

Digging Deep Into Deceit

How the deceit algorithm works. A deceit score is calculated for each individual Australian and International social media post, based in part on the LIWC dictionary developed by James W. Pennebaker (University of Austin, Texas) plus our own addition of selected emoticons that may reflect deceit. We have analysed 100,000’s of publicly available posts across Facebook, Twitter, Trip Advisor and Instagram to deliver the insights shared in part on this blog. Here is an infographic that explains the maths in fuller geeky detail. The algorithm is not infallible but is much better at detecting deceit than the average human who can spot only 54% of deceitful content. Many thanks to Suresh Sood from UTS for guiding and challenging us along the way.

11-TheAlgorithm_Content