Proposing a systematic approach for integrating traditional research methods into machine learning in text analytics in tourism and hospitality

This paper argues that the analysis of vast amounts of user-generated content, which are currently dominated by text analytics and machine learning, need more methodical incorporation of reliable traditional methodologies to facilitate deeper understanding of concepts and theory building. Specifically, a systematic approach that integrates machine learning and traditional research methods is needed to overcome inherent drawbacks of both approaches. A step-by-step methodological framework for the analysis of online reviews is proposed and demonstrated. An application of the framework with an example drawn from the context of understanding authenticity in dining experiences illustrates its usefulness in the investigation of complex concepts. This paper represents one of the first attempts to systematise an integrated learning approach to understand complex concepts and build theories in tourism and hospitality, contributing to more rigourous procedures for processing and analysing large data sets of user-generated content.