Machine learning for prediction of business company failure in hospitality sector

Abstract

The dynamic of success and failure in business is an open question. Since the seminal work of Altman, a considerable flow of literature has attempted to construct and test business failure models. More recently, the development of artificial intelligence has demonstrated great potential and several studies have shown that machine learning techniques achieved better performance than traditional statistical methods. We approach the prediction of company success/failure in a supervised learning framework, by building a dataset of financial indicators of companies for which it is known if they were successful or not. Our sample is with 34 financial indicators characterizing 1642 companies. We built multiple models based on this dataset to study their prediction performance on new previously unseen data from different companies. We measure each model’s prediction ability in terms of the usual classification metrics used by the machine learning research community, namely precision, recall, f-score and accuracy.