作者: Steven V. Viscido , Elizabeth E. Holmes
DOI: 10.1016/J.ENVSOFT.2010.05.010
关键词: Simulated annealing 、 Time series 、 Artificial intelligence 、 Search algorithm 、 Statistical model 、 Coding (social sciences) 、 Model selection 、 Machine learning 、 Software 、 Computation 、 Computer science
摘要: Understanding species interactions is critical to discovering community dynamics. Recently, statistical methods for estimating interaction strengths from time series data have been developed based on multivariate auto-regressive first-order, or MAR(1), models. However, the complex coding required presents a substantial barrier most ecologists. We LAMBDA, software program that allows users easily fit MAR(1) models multi-species data. The LAMBDA package covers: input and transformation, selection of include via search algorithm model selection, estimation parameters conditional least squares (CLS) regression two different maximum-likelihood (ML) algorithms, confidence intervals bootstrapping, computation stability properties using estimated model. describe performance tests variability estimates, speed, CLS versus ML simulated