Double weighted methodology: A weighted ensemble approach to handle concept drift in data streams

Data Streams are instances that arrive at a very rapid rate with changes in underlying conceptual distributions. Many ensemble learning approaches were developed to handle these changes in the dataset, which proved to be better than a single classifier system. In our work, we will discuss the framework of our new approach, Double Weighted Methodology and empirically prove it to be better than the existing single classifier approaches and the online ensemble approaches. Empirical results would prove that our approach is highly competitive, giving good accuracy and speed in handling and identifying drifts in data, irrespective of noise present in the dataset.