This study proposes a novel autoencoder-like semi-nonnegative matrix factorization (NMF) multiple clustering (ASNMFMC) model that generates multiple non-redundant, high-quality clustering. The nonnegative prope
AEs are instrumental in image classification. Methods like Semi-supervised stacked distance autoencoder (Hou et al.2020) enhance feature representation by incorporating semi-supervised learning, utilizing both labeled and unlabeled data to learn inter-data point distances. Deep Convolutional Autoencoders (...
NMF was implemented with the number of bases selected from validation data. Using DNN, the topology of 513–150–150–150–1026 with three 150-unit hidden layers was employed. In the implementation of DRNN, two recurrent hidden layers were constructed in a topology of 513–150–150–1026 (...
Second, conventional NMF models do not impose linear independence constraints, which leads to multiple potential factorization solutions to a given matrix. Differences between these solutions can be large, which incurs the additional task of identifying the optimal solution to all feasible ones. This ...