Interpretable Business Survival Prediction

TL;DR

This paper develops interpretable classifiers for business survival prediction using extensive feature engineering on geographic, mobility, attribute, and linguistic data from social networks, highlighting the importance of qualitative features.

Contribution

It introduces a comprehensive feature engineering approach and an interpretability framework to improve and explain business survival prediction models.

Findings

Qualitative features like business attributes and linguistic data have high predictive power.

The models achieved AUC scores of 0.72 and 0.67 for different feature sets.

Interpretability explanations can identify reasons for business survival from review texts.

Abstract

The survival of a business is undeniably pertinent to its success. A key factor contributing to its continuity depends on its customers. The surge of location-based social networks such as Yelp, Dianping, and Foursquare has paved the way for leveraging user-generated content on these platforms to predict business survival. Prior works in this area have developed several quantitative features to capture geography and user mobility among businesses. However, the development of qualitative features is minimal. In this work, we thus perform extensive feature engineering across four feature sets, namely, geography, user mobility, business attributes, and linguistic modelling to develop classifiers for business survival prediction. We additionally employ an interpretability framework to generate explanations and qualitatively assess the classifiers' predictions. Experimentation among the feature sets reveals that qualitative features including business attributes and linguistic features have the highest predictive power, achieving AUC scores of 0.72 and 0.67, respectively. Furthermore, the explanations generated by the interpretability framework demonstrate that these models can potentially identify the reasons from review texts for the survival of a business.