Fast semi-supervised classification based on bisecting clustering
Xiaolan Liu, Jingao Liu, Zhifeng Hao, Zhiyong Lin · 2010
In this paper, we propose a fast semi-supervised learning algorithm based on the bisecting clustering. The key idea of the proposed algorithm is dividing data into two sub clusters each time by using bisecting clustering and parts of the features of the data. The time complexity of the algorithm is nearly linear to the data size. Numerical comparisons with several existing methods for the UCI datasets and benchmark datasets verify the effectiveness of our method.