The Random Forest Kernel and other kernels for big data from random partitions

We present Random Partition Kernels, a new class of kernels derived by demonstrating a natural connection between random partitions of objects and kernels between those objects. We show how the construction can be used to create kernels from methods that would not normally be viewed as random partitions, such as Random Forest. To demonstrate the potential of this method, we propose two new kernels, the Random Forest Kernel and the Fast Cluster Kernel, and show that these kernels consistently outperform standard kernels on problems involving real-world datasets. Finally, we show how the form of these kernels lend themselves to a natural approximation that is appropriate for certain big data problems, allowing $O(N)$ inference in methods such as Gaussian Processes, Support Vector Machines and Kernel PCA.

Assessing effects of site characteristics on remediation secondary life cycle impact with a generalised framework

The 'sustainable remediation' concept has been broadly embraced by industry and governments in recent years in both the US and Europe. However, there is a strong need for more research to enhance its 'practicability'. In an attempt to fill this research gap, this study developed a generalised framework for selecting the most environmentally sustainable remedial technology under various site conditions. Four remediation technologies were evaluated: pump and treat (P&T), enhanced in situ bioremediation (EIB), permeable reactive barrier (PRB), and in situ chemical reduction (ISCR). Within the developed framework and examined site condition ranges, our results indicate that site characteristics have a profound effect on the life cycle impact of various remedial alternatives, thus providing insights and valuable information for determining what is considered the most desired remedy from an environmental sustainability perspective. © 2014 © 2014 University of Newcastle upon Tyne.