A text clustering method based on LSASOM is proposed. The text eigenvector is represented by Latent Semantic
Analysis (LSA), which embodies the semantic relation of the eigen words, and realizes the dimension reduction of the eigenvector.
SOM network algorithm is used to nonsupervised selforganizing study, and the weight vectors of neural network are continuously
adjusted in order to achieving the cluster targets. It requires no predefined number of clusters and can create a new species of text in
any right position, and remedies the defect of traditional methods.