This is an example of a HTML caption with a link.


13th International Conference on Mobile Data Management. July 23-26, 2012, Bengaluru, India.

Mobile Data Stream Mining: From Algorithms to Applications

Printer-friendly versionPDF version

Mobile devices are increasingly becoming the central computing and communication device in people’s lives. Devices today are equipped with a growing number of sophisticated embedded sensors such as an accelerometer, digital compass, gyroscope, GPS, microphone, light intensity sensor, and camera. This creates the opportunity to develop applications that leverage on the sensing capability of these mobile devices, as well as data from other sensors such as bio/body sensors.


Mobile data stream mining is a key technology for real-time analysis of data streams generated on-board the phone itself for the data generated by sensors on the phone and/or in close proximity to the phone. The significant advantages that mobile data stream mining provides over traditional strategies for leveraging the phone as a “transmission device” for sensor data, are as follows: reduce the amount of data transmitted from the phone to servers/the cloud, as well as reduce the energy/battery usage on the phone due to transmission of sensor data. Mobile data stream mining is particularly significant for applications that need real-time analysis of continuous data streams such as such as mobile crowd sensing, mobile activity recognition, intelligent transportation systems, mobile healthcare, and so on.


Mobile data stream mining techniques and algorithms typically focuses on adapting data stream mining techniques to be operational in the context of mobile devices. As such, the aim is to enable data stream mining to be performed in a manner such that it is congruent with the limited computational resources, screen real-estate, and energy considerations of the mobile device. Adaptive strategies that focus on adaptation of data rates and fine tuning of processing parameters have been shown to significantly enhance the longevity of continuous real-time processing of data mining in mobile environments. Adaptation can enable, if not guarantee, the continuity, cost-efficiency and consistency of a ubiquitous data stream mining application. In this context generic adaptation strategies termed Granularity-based Adaptation that can be used with any data stream mining technique running on a resource-constrained device have been developed. These approaches facilitates adaptation of data stream mining algorithms to varying data rates and available computational resources in mobile devices by innovative strategies to perform knowledge integration, controlling the rate of stream mining and varying the accuracy levels of the discovered patterns. In addition to such adaptation techniques for mobile data stream mining, numerous “light-weight” mining algorithms for varies types of analysis such as clustering, classification, concept drift detection, change detection, frequent items analysis and so have been developed. Finally, a number of applications/systems for mobile data stream mining have also been presented in the literature including: Mobimine, VEDAS, and MOLEC. Finally, an integrated toolkit for mobile data stream mining – Open Mobile Miner - has been developed to facilitate rapid deployment of applications in this area. There is also an emerging focus on visualization techniques and strategies for mobile data mining. In summary, while the last few years have seen continuous evolution of research in this area, mobile data mining has now finally come of age with significant wider interest in this domain.


This tutorial will provide a first principals introduction to data stream mining. This will be followed by a detailed review of adaptation strategies for mobile data mining and mobile data stream mining algorithms. The final segment of the tutorial will present the Open Mobile Miner toolkit and a number of application case studies that leverage mobile data stream mining. 



Speaker Bios:

Dr. Shonali Krishnaswamy

Institute for Infocomm Research (Singapore) and Monash University (Austalia)

Email: and

Shonali Krishnaswamy is the Deputy Head of the Data Mining Department at the Institute for Infocomm Research (I2R), Singapore and an Associate Professor in the Faculty of Information Technology at Monash University, Australia. Her research interests include data stream mining in mobile/embedded environments, and distributed and large-scale data mining, with an emerging interest in mobile crowd sensing and mobile activity recognition. In 2011, Shonali was part of all the teams that won all the four challenge tasks in the EU OPPORTUNITY Activity Recognition Challenge 2011. She has authored over 150 research publications and has been of the following awards: Monash University Vice-Chancellors Award for Excellence in Research by an Early Career Researcher, IBM Innovation Award (UIMA), Faculty of Information Technology Early Career Researcher Award and an Australian Post-Doctoral Fellowship from the Australian Research Council.



Prof. Joao Gama

Associate Professor

Laboratory of Artificial Intelligence and Computer Science

University of Porto, Portugal


João Gama is Associate professor at the University of Porto and researcher at LIAAD-Inesc Tec. He served as PI in several FCT projects in learning adaptive systems. He published more than 110 papers in major International conferences and journals, served as PC chair at ECML05, DS09, ADMA09, and Conference Chair at IDA11.He co-organized a series of workshops on learning from data streams in conjunction with ECML-PKDD, KDD, SAC and ICML. He is member of the editorial board of the Machine Learning Journal, DAMI, NGC, and PAI and he is author of a recent book in Knowledge Discovery from Data Streams.


Dr Mohamed Medhat Gaber

Senior Lecturer

University of Portsmouth, United Kingdom


Dr Mohamed Medhat Gaber is a Senior Lecturer at the University of Portsmouth, UK. He received his PhD from Monash University, Australia in 2006. He then held appointments with the University of Sydney, CSIRO, and Monash University, all in Australia. He has published more than 80 papers and edited/co-edited 4 books on data mining and knowledge discovery. Mohamed has served in the program committees of major conferences related to data mining, including ICDM, PAKDD, ECML/PKDD and ICML. He has also co-chaired over 10 workshops and special sessions on various data mining topics. Dr Gaber is recognised as a fellow of the UK Higher Education Academy (HEA). He is also a member of the International Panel of Expert Advisers for the Australasian Data Mining Conferences. In 2007, he was awarded the CSIRO teamwork award. 

Important Dates (more...)
  • Abstract Submission: December 12, 2011, 23:59 PT
  • Paper Submission: December 12, 2011, 23:59 PT
  • Acceptance Noification: February 20, 2012
  • Camera-ready due: April 27, 2012
  • Conference: July 23 to July 26, 2012
Latest Tweets (more...)