Ordinal regression with explainable distance metric learning based on ordered sequences

The purpose of this paper is to introduce a new distance metric learning algorithm for ordinal regression. Ordinal regression addresses the problem predicting classes which there natural ordering, but real distances between are unknown. Since walks fine line standard and classification, it common pitfall either apply regression-like numerical treatment variables or underrate information applying nominal classification techniques. On different note, discipline that has proven be very useful when improving distance-based algorithms such as nearest neighbors classifier. In addition, an appropriate can enhance explainability model. our study we propose approach distance, called chain maximizing learning. It based on maximization ordered sequences in local neighborhoods data. This takes into account all data without making use any two extremes regression, able adapt class separations not clear. We also show how extend learn non-linear setup using kernel functions. have tested several problems, showing high performance under usual evaluation metrics domain. Results verified through Bayesian non-parametric testing. Finally, explore capabilities terms case-based reasoning approach. these empirically datasets, experiencing significant improvements over with traditional Euclidean neighbors.