In practical applications, data stream classification faces significant challenges, such as high cost of labeling instances and potential concept drifting. We present a new online active learning ensemble framework for drifting data streams based on a hybrid labeling strategy that includes the following: 1) an ensemble classifier, which consists of a long-term stable classifier and multiple dynamic classifiers (a multilevel sliding window model is used to create and update the dynamic classifiers to effectively process both the gradual drift type and sudden drift type data stream) and 2) active learning, which takes a nonfixed labeling budget, supports on-demand request labeling, and adopts an uncertainty strategy and random strategy to label instances. The decision threshold of the uncertainty strategy is adjusted dynamically, i.e., when concept drift occurs, the threshold is gradually reduced to query the most uncertain instances in priority to reduce the request expense as much as possible. Experiments on synthetic and real data sets show that precise prediction accuracy can be obtained by the proposed method without increasing the total cost of labeling, and that the labeling cost can be dynamically allocated according to the concept drift.

Existing clustering algorithms are weak in extracting peak intervals for clustering time series data. In this paper, we propose a new clustering algorithm named Peak Interval Extraction Strategy based Hierarchical Clustering method (PIES_HC) for clustering time series data. The proposed PIES_HC algorithm can effectively exploit inherent peak interval information of a time series data set to enhance clustering performance. The main contributions of our work include the design of a strategy to extract peak intervals of time series data and the development of a new similarity measure method based on peak intervals. With a synthetic data set and three real-life data sets, our experimental results confirm that the proposed PIES_HC algorithm outperforms time series clustering algorithms based on Euclidean Distance and Dynamic Time Warping in terms of Accuracy.

In learning to classify data streams, it is impractical and expensive to label all of the instances. Online active learning over streaming data poses additional challenges for its increasing volumes and concept drifts. We propose a new online paired ensemble active learning framework consisting of a stable classifier and a timely substituted dynamic classifier to react to different types of concept drifts. Classifiers are built in block based way and will learn new instances incrementally online. According to a combination strategy of uncertainty strategy and random strategy, the decision whether to label the incoming instance for the updating of the stable classifier and the dynamic classifier will be made. Experimental evaluation results on real datasets show the advantage of the proposed work in comparison with other approaches.

In the domain of multi-label data and big data, it is hard for most classification methods to generate explainable classify rules, which makes it harder to get impact factors for classification. In this paper, contrast pattern equivalence class was discussed and mined to determine main impact factors for different classes, considering combination of frequent pattern mining and data classification. The rationality and feasibility to determine classification impact factors through contrast pattern equivalence classes mining was theoretically analyzed. Also the mine process and result analysis was shown through an experiment on the real dataset. Experimental results show the validity of the method proposed.