Updated SVM data and methods

In this release we have updated the SVM results. We used R to run a support-vector regression on the ABM data, and achieved much better results on this attempt than in our previous efforts. We also added new comparison plots for these results, and new multipanel comparisons to show the performance of every ML method for each scenario.

All the new SVM code, data and outputs are available in the 'SVM in R' folder. New multipanel comparison graphs are in the Comparison Plots folder.

v1.0 -- First release of ABM emulation tests

These Mathematica notebooks demonstrate our testing of various machine-learning methods and their ability to emulate the results of a complex agent-based model. These notebooks are linked to our forthcoming paper on the topic, which will be linked in the Readme once available.

This work will be expanded over time to demonstrate how to effectively use these methods to emulate a complex ABM, and thereby drastically shorten the time required to perform detailed analyses of that model by producing a surrogate model that runs much more quickly.