Predicting health inspection results from online restaurant reviews

TL;DR

This paper demonstrates that linguistic features extracted from online restaurant reviews can accurately predict official health inspection results with over 90% accuracy, leveraging simple classification methods.

Contribution

It introduces a novel approach of using linguistic analytics on online reviews to predict health inspection outcomes, combining keyword and topic features.

Findings

Achieved over 90% prediction accuracy.

Linguistic features effectively predict health inspections.

Simple SVM classifiers are sufficient for high accuracy.

Abstract

Informatics around public health are increasingly shifting from the professional to the public spheres. In this work, we apply linguistic analytics to restaurant reviews, from Yelp, in order to automatically predict official health inspection reports. We consider two types of feature sets, i.e., keyword detection and topic model features, and use these in several classification methods. Our empirical analysis shows that these extracted features can predict public health inspection reports with over 90% accuracy using simple support vector machines.