Abstract:

This paper presents a new unsupervised feature extraction method based on the obvious quantum entangled model and the latent quantum co-occurrence model to solve the problems that current text clustering methods don’t support incremental clustering and distributed computing, which is the foundation for the text clustering in Internet environment and single and multitext summary. The model without the support of domain knowledge maintains a good information clustering effect after moving ca 96% of the redundant features.Theory analysis and numerical experiments show that this model is effective.

Key words: unsupervised, feature selection, entangling model, window function