Hi GeoHero, Thanks for the clip. It was really insightful. I have recently been looking into contents in this area myself. I think 95% accuracy is relatively high. I came across a study recently that did a comparative analysis of traditional classification methods and the SVM classification accuracies. The SVM did come out tops, so I reckon most machine learning algorithms will. I however wonder how the 1% false positive metric compares with traditional methods. Any ideas? Please, I would also like to know if the source code for this machine learning method is open source (available for free use), and if yes, where I could get it. I'd like to implement it on datasets regarding forests in the sub-saharan parts of Africa, particularly my country Nigeria. Many thanks for the content again.