Efficient Concept Evolution Detection In Data Stream

In current years we have seen extraordinary growth in the application of data mining. Lot of work is done on infinite length, concept drift, and concept evolution. In that case we try to solve problem of continuous and fast data stream plus drift in . Previous work had done on the detection of unseen class. In this project we enhance the unseen class detection process with the additional method. Here we used flexible decision boundary and gini coefficient for batter result. In addition we are working on the simultaneous multiple unseen class detection process.