Figure 2 Results of conditional random forest model (a machine learning algorithm). Number of observations used to build the model were 84 and number of trees were 500 (left side). Predicted versus observed plot, showing accuracy of model (right side).