Exploiting evolving micro-clusters for data stream classification with emerging class detection

TL;DR: In this paper, the authors present three novel drift detection tests, whose test statistics are dynamically adapted to match the actual data at hand, based on a rank statistic on density estimates for a binary representation of the data, the second compares average margins of a linear classifier induced by the 1norm support vector machine (SVM), and the last one is based on the average zero-one, sigmoid or stepwise linear error rate of an SVM classifier.

TL;DR: A new online semi-supervised learning algorithm is proposed by modeling concept drifts with a set of micro-clusters that are dynamically maintained to capture the evolving concepts with error-based representative learning and yield high classification performance compared to many state-of-the-art algorithms.

TL;DR: In this paper , the authors present a comprehensive survey of techniques to deal with recurring changes in data streams and present open challenges and future research trends, as well as future directions on the usage of machine learning techniques for data streams.

TL;DR: In this article, the authors present a review of intrusion detection systems from the perspective of machine learning and present the three main challenges of an IDS, in general, and of IDS for the Internet of Things (IoT), in particular, namely concept drift, high dimensionality, and computational complexity.

TL;DR: A new experimental data stream framework for studying concept drift, and two new variants of Bagging: ADWIN Bagging and Adaptive-Size Hoeffding Tree (ASHT) Bagging are proposed.

TL;DR: In this paper, the authors present online versions of bagging and boosting that require only one pass through the training data, and compare the online and batch algorithms experimentally in terms of accuracy and running time.

TL;DR: An earthquake reporting system for use in Japan is developed and an algorithm to monitor tweets and to detect a target event is proposed, which produces a probabilistic spatiotemporal model for the target event that can find the center of the event location.

TL;DR: DDD maintains ensembles with different diversity levels and is able to attain better accuracy than other approaches, outperforming other drift handling approaches in terms of accuracy when there are false positive drift detections.