A Computational Approach to Automatic Prediction of Drunk Texting

TL;DR

This paper presents a computational method for automatically predicting whether a text was written under the influence of alcohol, using stylistic and N-gram features on labeled tweets.

Contribution

It introduces the task of drunk-texting prediction and demonstrates that textual signals can be effectively exploited for this purpose.

Findings

Text contains detectable signals of drunkenness

N-gram and stylistic features improve prediction accuracy

First quantitative evidence of drunk-texting detection from text

Abstract

Alcohol abuse may lead to unsociable behavior such as crime, drunk driving, or privacy leaks. We introduce automatic drunk-texting prediction as the task of identifying whether a text was written when under the influence of alcohol. We experiment with tweets labeled using hashtags as distant supervision. Our classifiers use a set of N-gram and stylistic features to detect drunk tweets. Our observations present the first quantitative evidence that text contains signals that can be exploited to detect drunk-texting.