Hof, Steve

Project title: Hidden Markov Models for Estimating Difficult to Measure Populations

Department: Mathematics and Statistics

Faculty supervisor: Dr. Laura Cowen

"Population sizes of small or otherwise ‘difficult to mark individually’ animals are notoriously difficult to estimate due to the necessity of ‘batch marking’ techniques which make estimating the likelihood of a population model difficult. In her research Dr. Cowen has demonstrated that Hidden Markov Models can provide a unified approach to population estimation whether the individuals of the population can be easily marked or not. This allows for the simultaneous estimation of population sizes, immigration and survival rates as well as efficient estimation of standard errors and model selection methods using standard likelihood techniques. In this project we hope to write and release an intuitive and efficient R package allowing for the easy implementation of these methods by fellow researchers or other individuals not necessarily intimately familiar with the underlying math."