Handling the varying classes of data in news feed

Computer Engineering

Clustering of text data stream has a huge emphasis on data mining. Clustering of text stream used in news group filtering, document organization and category definition and clustering etc. Concept drift is the underlying distribution of the data that is varying. Previous systems have used live concept drifting detection methods which have used error rate classification. To overcome the drawbacks of those previous methods we have new efficient clustering method using three layer concept drift in text data streams. This new method adapt to real time changes rapidly and rigorously. In this, three layer indicate the layer of label space, the layer of feature space and the layer of mapping relationships between labels and features respectively.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.