Java Deep Learning Projects

Deep learning for clustering analysis

Clustering analysis is one of the most widely used data-driven task. To date, existing clustering analysis techniques use classical clustering algorithms such as k-means, bisecting k-means, or the Gaussian mixture model. In particular, the k-means clustering algorithm and its several variants have been proposed to address issues with higher-dimensional input spaces.

However, they are fundamentally limited to linear embedding. Hence, they cannot model nonlinear relationships. Nevertheless, fine-tuning in these approaches is based on only cluster assignment hardening loss. Therefore, a fine-grained clustering accuracy cannot be achieved. 

In short, relatively less research has focused on deep-learning-based ...