Adaptive Quantum Clustering and Its Incremental Training

Abstract

An Adaptive Quantum Clustering algorithm (AQC) and its incremental training approach are proposed. AQC employs Schrödinger Equation to find core data. The equation is equipped with density estimation and a new metric that are learned from support vector clustering process. Based on core data, the grouping method of AQC assigns core data into clusters. AQC is trained incrementally through selecting valuable data from coming batch. Experiments suggest AQC improves clustering accuracy and efficiency over its counterparts, and achieves competitive performance with some state of the arts.