The new method establishes vector space model of term weight according to the theory of latent semantic index, and may eliminate disadvantageous factors.

• 该方法应用lsi理论来建立文本集的向量空间模型；在词条的权重中引入了语义关系；消减了原词条矩阵中包含的"噪声"因素；从而更加突出了词和文本之间的语义关系.