Predicting distribution of a new forest disease using one-class SVMs

In California, a newly discovered virulent pathogen (Phytophthora ramorum) has killed thousands of native oak trees. Mapping the potential distribution of the pathogen is essential for decision makers to assess the risk of the pathogen and aid in preventing its further spread. Most methods used to map potential ranges of species (e.g. multivariate or logistic regression) require both presence and absence data, the latter of which is not always feasibly collected. We present the one-class support vector machine (SVM) to predict the potential distribution of sudden oak death in California. The model was developed using presence data collected throughout the state, and tested for accuracy using a 5-fold cross-validation approach. The model performed well, and provided 91% predicted accuracy. We believe one-class SVM when coupled with geographical information systems (GIS) become a very useful method to deal with presence-only data in ecological analysis over a range of scales.