Abstract:
Pattern recognition, data mining, and machine learning all face clustering issues. Traditional shallow-structured clustering algorithms cannot uncover the interdependence of complex data features in latent space. Since they can learn promising latent representations from original data, deep generative models like autoencoder (AE), variational AE (VAE), and generative adversarial network (GAN) have had great success in many unsupervised applications. First, we propose a Wasserstein GAN with gradient penalty (WGAN-GP) and VAE with a Gaussian mixture prior-based clustering method. The WGAN-GP generator is formulated by drawing samples from VAE’s probabilistic decoder. A Student’s-t mixture prior-based variant of the proposed deep generative model improves clustering and generation when outliers are present. Clustering analysis and sample generation experiments validate our deep generative models. The proposed approach can train the model more stable, improve clustering accuracy, and generate realistic samples compared to other deep generative model-based clustering approaches.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
