Binary Information Press
There are some large mixed type datasets that sample number is very large and attribute space tends to be stationary. In order to cluster large mixed type dataset efficiently and effectively, a novel generic clustering framework based quasi-hyper image mapping is proposed. The large dataset is mapped on quasi-hyper-image to get a small quasi-hyper image pixel dataset. Because clustering the small quasi-hyper image pixel dataset is rapid, the large dataset can be clustered rapidly on the base of dataset mapping.