A new program can detect if someone is drunk by reading their text messages.
A team from India's IIT-Bombay and Monash University, Melbourne developed the program, which uses text-based analysis to determine if someone is under the influence of alcohol.
"This is a first-of-its-kind work that provides quantitative evidence that a text-based analysis may be useful for drunk-texting prediction," says Aditya Joshi, a Ph.D. student from IITB-Monash Research Academy.
The team used a statistical classifier for prediction, with two sets of features, "N-gram"-based and "stylistic," which qualify typical styles of writing that drunk texts may have, such as high sentiment-bearing words, capitalization, and spelling mistakes.
In a test involving tweets, the team obtained a labeled dataset by using hashtags that people add to tweets. As a result, a hashtag can reveal whether someone is drunk at the time of writing their tweet. The team's algorithm predicted 64 percent of the tweets written under the influence of alcohol.
"The drunk-texting prediction system is an automatic tool that will run constantly in the background and allow, say, police officers, to identify who is sending drunk texts in their locality," says IIT-Bombay's Abhijit Mishra.
From Indian Express
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA