Export Reviews, Discussions, Author Feedback and Meta-Reviews

An initial-visit emitting random walk is a random walk that emits an output only when visiting a state for the first time. Previous researchers have used these walks to model human behavior data on memory search tasks (e.g., name as many animals as you can). The contributions of this paper are that it proposes a new and efficient way of computing a likelihood function for the model, it proves that the resulting maximum likelihood estimates are consistent, it presents a way of estimating parameter values using a regularized likelihood function, and it shows, via computer simulation, that the proposed parameter estimation procedure performs better than alternatives on a toy task and on real-world data collected (by others) in experiments with people. I should start by stating that this paper relies on areas of mathematics that I know nothing about. That said, I have a mixed opinion of this paper.