Improved Logistic Regression Approach to Predict the Potential Distribution of Invasive Species Using Information Theory and Frequency Statistics

Abstract

The Predictive models of the potential distribution of invasive species are important for managing the growing invasive species crises. However, for most species absence data are not available. Presented with the challenge of developing a model based on presence-only information, we developed an improved logistic regression approach using Information Theory and Frequency Statistics to produce a relative suitability map. Logistic regression model selection was based on Akaike?s Information Criterion (AIC). Based on the weighted average model we provided the quantile statistics method to compartmentalize the relative habitat-suitability in native ranges. Finally, we used the model and the compartmentalize criterion developed in native ranges to project onto exotic ranges to predict the invasive species? potential distribution.