Cumulative link models for deep ordinal classification
This work addresses ordinal classification tasks, such as rating or ranking problems, by enhancing deep learning models with statistical ordinal techniques, representing an incremental improvement over existing approaches.
The paper tackled ordinal classification by proposing a deep convolutional neural network that uses cumulative link functions and a weighted Kappa-based loss function, resulting in improved performance over nominal models and other robust methods in experiments on two ordinal problems.
This paper proposes a deep convolutional neural network model for ordinal regression by considering a family of probabilistic ordinal link functions in the output layer. The link functions are those used for cumulative link models, which are traditional statistical linear models based on projecting each pattern into a 1-dimensional space. A set of ordered thresholds splits this space into the different classes of the problem. In our case, the projections are estimated by a non-linear deep neural network. To further improve the results, we combine these ordinal models with a loss function that takes into account the distance between the categories, based on the weighted Kappa index. Three different link functions are studied in the experimental study, and the results are contrasted with statistical analysis. The experiments run over two different ordinal classification problems and the statistical tests confirm that these models improve the results of a nominal model and outperform other robust proposals considered in the literature.