Incremental Learning Algorithms for Fast Classification in Data Stream

Classification is one of the most commonly used data mining methods which can make a prediction by modeling from the known data. However, in traditional classification, we need to acquire the whole dataset and then build a training model which may take a lot of time and resource consumption. Another drawback of the traditional classification is that it cannot process the dataset timely and efficiently, especially for real- time data stream or big data. In this paper, we evaluate a lightweight method based on incremental learning algorithms for fast classification. We use this method to do outlier detection via several popular incremental learning algorithms, like Decision Table, Naive Bayes, J48, VFI, KStar, etc.