TCBLSA: a new method of text clustering

This paper presents a new method of text clustering based on the theory of latent semantic analysis (LSA) called TCBLSA method. The vector space model (VSM) of term weight is constructed by the theory of LSA and the TF.IDF method. The present method decreases the dimension of vector, and eliminates disadvantageous factors in the VSM. Furthermore, the method advances the speed and precision of text clustering. Through analyzing experimental data, we demonstrate that the TCBLSA method is effective and feasible for text clustering.