Text clustering method based on the iteration convergence of initial centers

In this paper according to the center of the text set and the initial cluster, a set of directions with good discriminations have been chosen to construct the IMIC coordinate. In this coordinate system re-scaling functions were constructed for each axis which are used to improve the effectiveness of the cluster policy. After several iterations, the IMIC algorithm finally converged to the final solution. The time complexity of the IMIC algorithm remains the same level as that of the K-means. Experimental results show that, the IMIC algorithm has better clustering quality.