Moonlight and shelter cause differential seed selection and removal by rodents

Various environmental factors may influence the foraging behaviour of seed dispersers which could ultimately affect the seed dispersal process. We examined whether moonlight levels and the presence or absence of rodentshelter affect rodentseedremoval (rate, handling time and time of removal) and seedselection (size and species) among seven oak species. The presence or absence of safe microhabitats was found to be more important than moonlight levels in the removal of seeds. Bright moonlight caused a different temporal distribution of seedremoval throughout the night but only affected the overall removal rates in open microhabitats. Seeds were removed more rapidly in open microhabitat (regardless of the moon phase), decreasing the time allocated to seed discrimination and translocation. Only in open microhabitats did increasing levels of moonlight decrease the time allocated to selection and removal of seeds. As a result, a more precise seedselection was made under shelter, owing to lower levels of predation risk. Rodent ranking preference for species was identical between full/new moon in shelter but not in open microhabitats. For all treatments, species selection by rodents was much stronger than size selection. Nevertheless, heavy seeds, which require more energy and time to be transported, were preferentially removed under shelter, where there is no time restriction to move the seeds. Our findings reveal that seedselection is safety dependent and, therefore, microhabitats in which seeds are located (sheltered versus exposed) and moonlight levels in open areas should be taken into account in rodent food selection studies.

An Interval-based Multiobjective Approach to Feature Subset Selection Using Joint Modeling of Objectives and Variables

This paper studies feature subset selection in classification using a multiobjective estimation of distribution algorithm. We consider six functions, namely area under ROC curve, sensitivity, specificity, precision, F1 measure and Brier score, for evaluation of feature subsets and as the objectives of the problem. One of the characteristics of these objective functions is the existence of noise in their values that should be appropriately handled during optimization. Our proposed algorithm consists of two major techniques which are specially designed for the feature subset selection problem. The first one is a solution ranking method based on interval values to handle the noise in the objectives of this problem. The second one is a model estimation method for learning a joint probabilistic model of objectives and variables which is used to generate new solutions and advance through the search space. To simplify model estimation, l1 regularized regression is used to select a subset of problem variables before model learning. The proposed algorithm is compared with a well-known ranking method for interval-valued objectives and a standard multiobjective genetic algorithm. Particularly, the effects of the two new techniques are experimentally investigated. The experimental results show that the proposed algorithm is able to obtain comparable or better performance on the tested datasets.