PIECEWISE:
  Piecewise Linear Technique in Support Vector Machines and Applications

 

Financing: Other Funding Agencies (OTHER)

Project reference Nr.: BIL12/11T
Start: 2013-05-01
End: 2015-04-30

Description:

In this research, the combination of SVMs and PWL will be investigated insightfully and the resulted method will be used in real world applications. The group lead by Prof. Johan A.K. Suykens in KU Leuven has been working on machine learning and achieved a series of interesting results  including the least squares support vector machine (LS-SVM), which has been widely applied and cited more than four thousand times. The group lead by Prof. Shuning Wang in Tsinghua University has been working on piecewise linear technique, and their recent results include PWL programming and the latest compact representation of PWL functions with universal representation capability. The group lead by Prof. Dexian Huang in Tsinghua University has been working on process control and has obtained remarkable economic benefit in the petroleum chemical industry. The collaboration of the KU Leuven and Tsinghua teams is promising to develop new SVM methods with PWL techniques, and several applications in process control can be expected. The potential research directions include PWL SVM classification, PWL SVM regression, PWL kernel methods, PWL programming


 

SMC people involved in the project: