The use of machine learning algorithms for the study of business profitability: a new approach based on preferences

In recent years, researchers in the Field of Artificial Intelligence have developed a learning technique, namely, preference learning, that is suitable to be used for economic analysis. The present research empirically tests one of these models, which consists of a combination of LACE and RFE algorithms. The problem of forecasting the profitability of Spanish companies upon the basis of a set of financial ratios is used as a benchmark. The model provides forecasted rankings, which are a kind of information that is more useful for the economic analysts than the forecasted class memberships that traditional machine learning techniques provide.