A new approach: semisupervised ordinal classification

Semisupervised learning is a type of machine learning technique that constructs a classifier by learning from a small collection of labeled samples and a large collection of unlabeled ones. Although some progress has been made in this research area, the existing semisupervised methods provide a nominal classification task. However, semisupervised learning for ordinal classification is yet to be explored. To bridge the gap, this study combines two concepts “semisupervised learning” and “ordinal classification” for the categorical class labels for the first time and introduces a new concept of “semisupervised ordinal classification”. This paper proposes a new algorithm for semisupervised learning that takes into account the relationships between the class labels, especially class orderings such as low, medium, and high. We also performed an extensive empirical study that involves 10 benchmark ordinal datasets with different quantities of labeled samples varying from 15% to 50% with an increment of 5%, aiming to evaluate the performance of our method by combining different base learners. The experimental results were also validated with a nonparametric statistical test. The experiments show that the proposed method improves the classification accuracy of the model compared to the existing semisupervised method on ordinal data.