A semi-supervised weighted clustering framework facing to hybrid attributes data streams

In order to solve weighted clustering analysis problem of infinite hybrid attributes data streams in finite space, it presents a weighted clustering and evolution analysis framework. This framework is based on decision-tree classify of little sample using a semi-supervised strategy. In order to record some necessary information of cluster group, it gives a definition of cluster feature vector group of hybrid attributes data streams. In order to update the feature weight group, it presents an adaptive optimization method of configuring feature weight group. It gives some necessary discuss about weighted clustering and evolution analysis framework, which is important to implement this framework. This framework can get better results sometimes.