A Recommendation Engine (RE) is an approach under machine learning, that presents personalized suggestions/references by predicting a user’s prospect favorites for a breadth of facilities. Concept Drift (CD) is a common problem in online supervised learning environments such as dynamic recommendation engines seen in e-commerce portals, where data changes over time. Although there are other CD detectors in the research that are now available, the supervised technique known as Adaptive Windowing 2 (ADWIN2) is the most recommended option for non-stationary, dynamic, and streaming data. Kalman Filters is known for the best object estimation in noisy and uncertain environments. The paper aims towards the concept of ADWIN2 approach for CD detection and working of Kalman Filters. This paper reviews and compares various types of techniques, other than Kalman Filter for CD detection with linear datasets. The paper ends with the detailed limitations of Kalman Filters and prominent application of Kalman Filter in Autonomous Vehicles (AVs) for the development of Smart Cities.