The performance of K-means and Gaussian mixture model (GMM) clustering depends on the initial guess of partitions. Typically, clus∗ corresponding author 1
Abstract— Ultra wide band (UWB) impulse radio (IR) technology calls for robust and low-complexity receiver techniques. State-of-the-art proposals are both coherent ML receivers, ...
This paper presents a general and efficient framework for probabilistic inference and learning from arbitrary uncertain information. It exploits the calculation properties of fini...
—In Dirichlet process (DP) mixture models, the number of components is implicitly determined by the sampling parameters of Dirichlet process. However, this kind of models usually...
Nowadays, the source localization has been widely applied for wireless sensor networks. The Gaussian mixture model has been adopted for maximum-likelihood (ML) source localization ...