type out semi - illegible tweets while you trip up around trying to encounter your Uber is the go - to end to a late night out in 2016 . However , those sloshed tweet could really be used for some adept , thanks to this machine learning - based algorithm .
A group of reckoner scientist from the University of Rochester has develop a car - learning algorithm that they ’ve trained to detect inebriated tweets . Their study was recently published in the journalarXiv .
The team analyzed over 11,000 geotagged tweet posted in New York City and Monroe County between July 2013 and July 2014 . From this natural selection , they filtered all the posts that mention alcohol - tie in buzzwords , which included “ drunk , ” “ tequila , ” “ beer , " “ hammered , ” and “ get consume . ” By give different value and “ free weight ” to each keyword , the computing machine can see if alcoholic drink - drinking is actually cite , while remaining cautious around deceptive words such as “ shoot , ” “ party , ” or “ nightspot , ” which might not necessarily be about drinking .
Using this composite plant of boozy data and further analysis of the words in the tweets , the machine is then able-bodied to decrypt whether or not the C. W. Post is about the Twitter - exploiter themselves being intoxicated , and then if the Twitter - exploiter was actually drinking at the clock time of twirp . As the tweet were geotagged , the computer was also able-bodied to find the location of where the tweeter had been toast .
“ We can dissect human mobility patterns ; we can study the relationship between demographic , neighborhood social system and health conditions in different zip codification , thus realize many aspect of urban life and environments , ” the researcher wrote . “ Research in these area and alcoholic drink use is chiefly based on surveys and census , which are costly and often obtain a holdup that hamper existent - time analytic thinking and response . Our results demonstrate that tweet can provide sinewy and fine - grained cues of activities going on in cities . ”
The computer scientist hope that this algorithm can be used as a instrument to accost alcohol consumption through public insurance policy , as well as provide a example for succeeding wellness - related nosecount and surveys .
( H / T : MIT Technology Review )
