Interactive spatial evolutionary computation based predictive modeling of rare plant species occurrences

In this paper, a habitat suitability model for rare plant species is developed by combining spatial predictive modeling techniques and an evolutionary computation method known as Interactive Evolution (IE). The interactive spatial evolutionary computation based predictive modeling enables an analyst to combine advanced mathematical geospatial and pattern recognition modeling techniques, conceptual knowledge, available empirical data, and expert-interactive dynamic data visualization techniques to predict species distributions. This methodology represents a step toward identifying regions where species are likely located to base argument for appropriate stratification in applying culturally relative conservation strategies to protect them and promote overall enhanced ecological integrity to the region. The model accuracy was assessed statistically by testing for independence between predicted occurrence and actual occurrence using cross-validation tests.