Enabling Lazy Learning for Uncertain Data Streams

Lazy learning concept is performing the k-nearest neighbor algorithm, Is used to classification and similarly to clustering of k-nearest neighbor algorithm both are based on Euclidean distance based algorithm. Lazy learning is more advantages for complex and dynamic learning on data streams. In this lazy learning process is consumes the high memory and low prediction Efficiency .this process is less support to the data stream applications. Lazy learning stores the trained data and the inductive process is different until a query is appears, In the data stream applications, the data records flow is continuously in huge volume of data and the prediction of class labels are need to be made in the timely manner. In this paper provide the systematic solution to overcome the memory and efficiency. In this paper proposed a indexing techniques it is dynamically maintained the historical or outdated data stream records. In this paper proposed the tree structure i.e. Novel lazy tree simply called Lazy tree or L-tree.it is the height balanced tree or performing the tree traversing techniques to maintain the trained data. These are help to reduce the memory consumption and prediction it also reduces the time complexity. L-tree is continuously absorb the newly coming stream records and discarded the historical. They are dynamically changes occurred in data streams efficiency for prediction. They are experiments on the real world data streams and uncertain data streams. In this paper experiment on the uncertain data streams .Our experimented uncertain data streams and real world data streams are obtained from UCI Repository.