Improving an ensemble of neural networks via a novel multi-class decomposition schema

The need for high recognition performance demands increasingly complex machine learning (ML) architectures, which might be extremely computationally burdensome to be implemented in real-world. This issue can be addressed by using an ensemble learning model to decompose the multi-class classification problem into many simpler binary classification problems, e.g. each binary classification problem can be handled via a simple multi-layer perceptron (MLP). The so-called one-versus-one (OVO) is a widely used multi-class decomposition schema in which each classifier is trained to distinguish between two classes. However, with an OVO schema each MLP is non-competent to classify instances of classes that have not been used to train it. This results in classification noise that may degrade the performance of the whole ensemble, especially when the number of classes grows. The proposed architecture employs a weighting mechanism to minimize the contribution of the non-competent MLPs and combine their outcomes to effectively solve the multi-class classification problem. In this work, the robustness to the classification noise introduced by non-competent MLPs is measured to assess in what conditions this translates in better classification accuracy. We test the proposed approach with five different benchmark data sets, outperforming both the baseline and one state-ofthe-art approach in multi-class decomposition algorithms.