Machine Learning With the Sugeno Integral: The Case of Binary Classification

In this article, we elaborate on the use of Sugeno integral in context machine learning. More specifically, propose a method for binary classification, which is used as an aggregation function that combines several local evaluations instance, pertaining to different features, or measurements, into single global evaluation. Due specific nature integral, approach especially suitable learning from ordinal data, i.e., when measurements are taken scales. This topic has not received much attention so far. The core problem itself consists identifying capacity underlying integral. To tackle problem, develop algorithm based linear programming. also includes technique transforming original feature values (local utility scores), well tuning threshold control flexibility classifier and mitigate overfitting training generalize our toward $k$ -maxitive capacities, where plays role hyperparameter learner. We present experimental studies, compare with competing approaches benchmark datasets.