Artificial Intelligence Structuration in Machine Learning

This research adopts a social construction perspective, utilizing Structuration Theory to explore the reciprocal interactions between data and algorithms. We propose an Artificial Intelligent Structuration framework for machine learning forecasting, which comprises three structural layers: the inference layer for the signification of data, the reference layer for the domination by algorithms, and the meta layer for the legitimation for insights. The proposed framework is demonstrated through a case study on hotel rate predictions. Empirical results indicate that integrating machine learning models within the meta layer significantly enhances prediction accuracy. This approach underscores the potential of structured interactions between data and algorithms to improve forecasting outcomes in business applications.