A visual multi-expert neural classifier
Chester J. Ornes, Jack Sklansky · Proceedings of International Conference on Neural Networks (ICNN'97) · 2002
When an automatic pattern classifier faces a difficult query, the user may benefit from knowing what data in the training set is illustrative of the query. We describe a high-performance neural network that in addition to classifying the query, allows the user to visualize the relationship between the query and the data in the training set. We show in applications, including medical diagnosis and image segmentation, that our classifier achieves low error rates while providing a visual explanation of classifier decisions. We demonstrate the properties of the classifier using synthetic data, and compare the visualization performance of the visual neural network to Kohonen's self-organizing map.