Gradient-Boosted-Regression-Trees

Gradient boosted regression trees were used to explore the results of simulating ASF spread in wild pigs across a landscape. Initial parameters varied (predictors) include:

• host density
• infectious period of living individuals
• infectious period of carcasses on landscape
• incubation period
• proportion that recover
• days to recovery

Outbreak traits of interest:

• landscape area affected at 1 year
• peak number of cases
• outbreak probability
• rate of spread
• global seroprevalence at 1 year

NOTE: If using these scripts, cross validate first and then run the regression code with the parameters resulting in the highest $R^2$ value. Additional scripts included to visualize results in both box and whisker plots and scatter plots as well as scripts used to process results of disease model.