Open challenges for data stream mining research

Abstract

Every day, huge volumes of sensory, transactional, and web data are continuously generated as streams, which need to be analyzed online as they arrive. Streaming data can be considered as one of the main sources of what is called big data. While predictive modeling for data streams and big data have received a lot of attention over the last decade, many research approaches are typically designed for well-behaved controlled problem settings, overlooking important challenges imposed by real-world applications. This article presents a discussion on eight open challenges for data stream mining. Our goal is to identify gaps between current research and meaningful applications, highlight open problems, and define new application-relevant research directions for data stream mining. The identified challenges cover the full cycle of knowledge discovery and involve such problems as: protecting data privacy, dealing with legacy systems, handling incomplete and delayed information, analysis of complex data, and evaluation of stream mining algorithms. The resulting analysis is illustrated by practical applications and provides general suggestions concerning lines of future research in data stream mining.

Improve IoT/M2M Data Organization Based on Stream Patterns

“…By using our generative stream model as a tailored feature to describe stream patterns we believe that our organization model will be further improved. It is worthwhile to mention that there are several academic works based on stream prediction and mining [16], but the same cannot be said about stream similarity and stream characterization. Further work needs to be done to assert some ideas expressed in this paper, but our stream characterization model appears to be a viable option.…”

“…While there are several academic works based on stream prediction and mining [16], the same can not be said about stream similarity. Most methods are based on longest common sub-sequence algorithm [17], [18].…”

Big Data Analysis for M2M Networks:Research Challenges and Open Research Issues

“…When applying machine-learning techniques for M2M communication unexpected problems may arise due to the complexity and processing time added by the scale of applications. The large scale makes trivial operations costly in terms of computation time and memory requirement and forces the system designers to reconsider the applied machine learning algorithms since the cost of learning arises primarily from bandwidth and disk reads [25,40]. In this section, a literature review is presented to discuss the challenges in data mining and machine learning for M2M applications.…”

Sliding Reservoir Approach for Delayed Labeling in Streaming Data Classification

“…A naive solution of the delayed labeling problem will be requesting labels immediately at the time of concept drift [15]. Then the framework waits for the labeling process to finish before building any updated models.…”