With the rapid development of information technology, TB-level or even PB-level data are continually generated from Internet systems, various cyber-physical systems, smart city networks, education systems and healthcare systems, scientific research systems, government departments and other aspects. These data formats vary from traditional static data to a dynamic data stream with the characteristics of being multi-source, heterogeneous, high-dimensional and dynamic. Such data can be termed as multi-source data stream. Mining of valuable information from the multi-source data stream is of significant importance.
Different from traditional data mining, multi-source data stream mining has several unique problems. As the data stream is continuously generated, intelligent mining algorithm should be performed on selected and appropriate window length data. Due to the different modals and dimensions, there is often a need to fuse the extracted features of multi-source data stream and preserve the complementarity of each modal data itself and with each other to eliminate redundancy. In addition, labeling of training samples is difficult because of the dynamics of the sample generation. In addition, the concept of drift detection and update mechanism are important aspects for multi-source data stream mining. Given these challenges, this Special Issue seeks to provide a novel mining model, features extraction and optimization mechanism, few shot learning methods, concept drift detection means and other intelligent algorithms and applications for multi-source data stream.
This Special Issue is aimed at encouraging researchers to present their latest work on intelligent mining algorithms for data stream.
Topics of interest include, but are not limited to:
- Machine learning for data stream mining;
- Features extraction and optimization for high-dimension data;
- Features fusion for network data;
- Few shot learning for data stream mining;
- Zero-shot learning for data stream mining;
- Deep learning for data stream mining;
- Transfer learning for data mining;
- Incremental learning for data stream mining;
- Concept drift detection for data stream mining;
- Real-world applications of data stream mining.
Submission deadline: 31 December 2022
Prof. Wei Wang, Tianjin Normal University, email@example.com
Prof. Tariq S. Durrani, University of Strathclyde, firstname.lastname@example.org
Prof. Xiantao Cai, Wuhan University, email@example.com
Prof. Qilian Liang, the University of Texas at Arlington, firstname.lastname@example.org
Before submitting your manuscript, please ensure you have carefully read the Instructions for Authors for EURASIP Journal on Advances in Signal Processing. The complete manuscript should be submitted through the EURASIP Journal on Advances in Signal Processing submission system. To ensure that you submit to the correct special issue please select the appropriate section in the drop-down menu upon submission. In addition, indicate within your cover letter that you wish your manuscript to be considered as part of the special issue. All submissions will undergo rigorous peer review and accepted articles will be published within the journal as a collection.
Sign up for article alerts to keep updated on articles published in EURASIP Journal on Advances in Signal Processing - including articles published in this special issue!