The major problem of online learning orincremental learning is that, target function is frequently changing over time. This problem iscommonly known as concept drift. Concept drift can beis further complicated if the dataset is class-imbalanced. There are different learning methods presented so far to handle concept drift like rule-based systems, decision trees, Naive Bayes, support vectormachines, instance based learning, ensemble ofclassifiers, etc. This learning method requires further tocombined with methods of drift detection in order toconstantly monitor the performance of concept drift,however online changes detection was failed. Inliterature there are many methods presented forlearning from data streams and drift detection, butmost of methods failed to achieve speed and accuracydue to data inconsistency. In this project our goal is to present efficient method for online and non-parametricdrift detection. This proposed method is based onrecently presented Hoeffdings Bounds and HDDM. Ithandles concept drift regardless of the learning modelto monitor the performance metrics measured duringthe learning process, to trigger drift signals when a significant variation has been detected. The existing system however as Naive Bayes classifier are havinglimitations, there is no scope to improve accuracy ofHDDM. The Propose system will be efficiently providedrift detection method for data stream mining toimprove accuracy.