Classification via Deep Fuzzy c-Means Clustering

Mojtaba Yeganejou, Scott Dick

Abstract: While deep learning has proven to be a powerful new tool for modeling and predicting a wide variety of complex phenomena, those models remain incomprehensible black boxes. This is a critical impediment to the widespread deployment of deep learning technology, as decades of research have found that users simply will not trust (i.e. make decisions based on) a model whose solutions cannot be explained. Fuzzy systems, on the other hand, are by design much more easily understood. We propose to create more comprehensible deep networks by hybridizing them with fuzzy logic. Our proposed architecture first employs a convolutional neural network as an automated feature extractor, and then clusters datasets in that feature space using the fuzzy c- means algorithm. After hardening the clusters, we employ Roccio's algorithm to classify the data points. Experiments on the MNIST dataset show that this classifier is comparable to other deep learning algorithms, and substantially more accurate than the same fuzzy classifier applied to the original image.