A likelihood approach for functional response models

作者：Okuyama Toshinori^*

来源：Biological Control, 2012, 60(2): 103-107.

DOI：10.1016/j.biocontrol.2011.10.008

• Maximum likelihood
• Sum of squares
• Predation

Functional response is an important determinant of community dynamics, and thus empirical methods for characterizing functional responses are as important in understanding ecological processes. The most commonly used method is based on the sum of squares, and the maximum likelihood method is rarely used. When the likelihood method is used, potentially inappropriate probability distributions such as binomial distributions are typically assumed for the number of prey eaten in experiments. In this study, I present a likelihood approach in which the probability distributions are generated by mechanistic understanding of predation processes using Monte Carlo simulations. An example is given on the Holling type II functional response model, but the method is flexible and allows characterization of a wide variety of functional response models. In the example, the likelihood method consistently resulted in superior estimates than the least squares method.

• 出版日期2012-2