detect concept drift in online data streams

To propose robust concept drift detection approach that balances between accuracy and time for detect concept drift (robust approaches that yield a balance between incremental learning and forgetting) this is my Requirements 1- basic approach to change detection in data streams uses two sliding windows over the data stream. 2- We want the guarantee that our test detects true changes with high probability, i.e.,it should have few false negatives camper with the basic references papers that shows in my problems. 4- Explain the effectiveness of the proposed approach 5- Design algorithm based on problem mentioned in pdf file 6- Explain algorithm 7- Explain how to make experiment 8- Experiment with MOA 9- Explain evaluation metric 10- validate the work 11- results in (plot drowning with matlap) and discusses the results 12- write the report 13- the solution cannot published before