The researchers used online gambling account data to develop models that could identify problem gambling. They used survey data and account data provided by two French authorities, the ARJEL and the FDJ. The researchers developed two models to predict problem gambling in skill-based and chance-based games, respectively. The first model produced good results and was best at identifying non-problem and problem gambling, but performed poorly at identifying low-risk and moderate-risk gambling. The second model produced moderate results, but was also better at identifying non-problem and problem gambling over low-risk and moderate-risk gambling.
