Anomaly Detection in Data Streams

Introduction

    Streaming data mining has become one of the most challenging research issues in the machine learning community. With the deployment of more and more IoT (Internet of Things) applications, a huge amount of data is generated as data streams in many real-world applications with fast rates. This leads to the so-called phenomenon of “data explosion”. Furthermore, due to huge volume of data streams and rapid rates of data generation, it is hardly possible to store all the data in the streams. This makes majority of if not all existing data mining methods facing challenges to handle streaming data.

    Anomaly detection is one of the most important problems in performance surveillance, contingency planning and decision making support. Anomaly detection in streaming data becomes even more important and challenging because in many cases, streaming data is not persistent and conventional machine learning algorithms are inherent with the problems of offline working principle, slow learning speed, lack of adaptive mechanism and over-dependency on manual intervention. This issue calls for innovation in streaming data processing, fast online machine learning algorithms for streaming data and handing uncertainty in streaming data.

    This special session aims to provide a forum for researchers and practitioners in the areas related to anomaly detection in steaming data to share their theoretical innovation and advancement of practical applications.

    This special session aligns with two themes of ISKE symposium -- Knowledge Engineering and Management and Practical Applications and Systems. It covers Data Mining and Knowledge Discovery, Pattern Recognition, Decision Support Systems, Service Computing and Mobile Computing, Business Intelligence, Intelligent Systems in smart cities.

Scope and Topics

The topics of interest include, but are not limited to:
  • Online real-time unsupervised learning and clustering for data streams
  • Online real-time supervised classification and regression for large data streams
  • Appropriate handling of data uncertainty in learning from large data streams
  • Tools and techniques for data stream mining in uncertain environments
  • Techniques to address drifts and shifts in data streams
  • Feature selection and extraction techniques for large data streams
  • Practical applications of anomaly detection techniques in large data streams
  • Applications of intelligent data processing techniques in large data streams

Special Session Organisers

  • Dr. Mukesh Prasad, Faculty of Engineering and Information Technology, Centre for Artificial Intelligence, University of Technology Sydney, E-mail: Mukesh.Prasad@uts.edu.au
  • Dr. Haiyan Lu, Faculty of Engineering and Information Technology, Centre for Artificial Intelligence, University of Technology Sydney, E-mail: Haiyan.lu@uts.edu.au
  • Dr. Sudhanshu Joshi, Faculty of Engineering and Information Technology, Centre for Artificial Intelligence, University of Technology Sydney, E-mail: sudhanshujoshi@doonuniversity.ac.in

Submission

Papers reporting original research results and experience are solicited. Each paper should have no more than 8 pages, including references and Illustrations, written in IEEE conference proceedings format (See the templates available at: Paper Templates). Papers must be written in English and submitted electronically as a PDF/Word file at EasyChair: https://easychair.org/conferences/?conf=iske2017, and please select the Anomaly Detection in Data Streams. Submission of a paper should be regarded as an undertaking that, if the paper is accepted, at least one of the authors must attend the conference to present the paper.

Important Dates


1 Paper Submission Deadline: Aug 1, 2017
2 Notification Acceptance: Sep 1, 2017
3 Camera-ready Paper and Registration: Sep 25, 2017