Harmonic mean similarity based quantum annealing for k-means
Jo Takano, Toshiaki Omori · Procedia Computer Science · 2018
Clustering is an important machine learning approach in analyzing big data. In this study, we propose a clustering method using quantum annealing (QA) based on harmonic average of purity and inverse purity. By using harmonic average of purity and inverse purity, we introduce the effect of quantum noise for clustering algorithm while considering both quality of clusters and quality of categories. Using benchmark data, we show the effectiveness of the proposed QA clustering based on the harmonic average of purity and inverse purity.