Fitting Multiple Models via Density Analysis in Tanimoto Space

Author(s):  
Luca Magri ◽  
Andrea Fusiello
2020 ◽  
Vol 635 ◽  
pp. 187-202
Author(s):  
T Brough ◽  
W Rayment ◽  
E Slooten ◽  
S Dawson

Many species of marine predators display defined hotspots in their distribution, although the reasons why this happens are not well understood in some species. Understanding whether hotspots are used for certain behaviours provides insights into the importance of these areas for the predators’ ecology and population viability. In this study, we investigated the spatiotemporal distribution of foraging behaviour in Hector’s dolphin Cephalorhynchus hectori, a small, endangered species from New Zealand. Passive acoustic monitoring of foraging ‘buzzes’ was carried out at 4 hotspots and 6 lower-use, ‘reference areas’, chosen randomly based on a previous density analysis of visual sightings. The distribution of buzzes was modelled among spatial locations and on 3 temporal scales (season, time of day, tidal state) with generalised additive mixed models using 82000 h of monitoring data. Foraging rates were significantly influenced by all 3 temporal effects, with substantial variation in the importance and nature of each effect among locations. The complexity of the temporal effects on foraging is likely due to the patchy nature of prey distributions and shows how foraging is highly variable at fine scales. Foraging rates were highest at the hotspots, suggesting that feeding opportunities shape fine-scale distribution in Hector’s dolphin. Foraging can be disrupted by anthropogenic influences. Thus, information from this study can be used to manage threats to this vital behaviour in the locations and at the times where it is most prevalent.


2020 ◽  
Vol E103.C (11) ◽  
pp. 588-596
Author(s):  
Masamune NOMURA ◽  
Yuki NAKAMURA ◽  
Hiroo TARAO ◽  
Amane TAKEI

2019 ◽  
Author(s):  
Theodosios Famprikis ◽  
James Dawson ◽  
François Fauth ◽  
Emmanuelle Suard ◽  
Benoit Fleutot ◽  
...  

<div> <p>Solid electrolytes are crucial for next‑generation solid‑state batteries and Na<sub>3</sub>PS<sub>4</sub> is one of the most promising Na<sup>+</sup> conductors for such applications. At present, two phases of Na<sub>3</sub>PS<sub>4</sub> have been identified and it had been thought to melt above 500 °C. In contrast, we show that it remains solid above this temperature and transforms into a third polymorph, γ, exhibiting superionic behavior. We propose an orthorhombic crystal structure for γ‑Na<sub>3</sub>PS<sub>4</sub> based on scattering density analysis of diffraction data and density functional theory calculations. We show that the Na<sup>+</sup> superionic behavior is associated with rotational motion of the thiophosphate polyanions pointing to a rotor phase, based on <i>ab initio</i> molecular dynamics simulations and supported by high‑temperature synchrotron and neutron diffraction, thermal analysis and impedance spectroscopy. These findings are of importance for the development of new polyanion‑based solid electrolytes.</p> </div>


2018 ◽  
Vol 6 (4) ◽  
pp. 424-428 ◽  
Author(s):  
Mayur D. Chaudhari ◽  
◽  
◽  
Archana S. Ghotkar

2010 ◽  
Vol 36 (9) ◽  
pp. 1295-1304 ◽  
Author(s):  
Yi-Hui ZHENG ◽  
Xin WANG ◽  
Shao-Yuan LI ◽  
Jian-Guo JIANG

2019 ◽  
Vol 13 (4) ◽  
pp. 325-333
Author(s):  
Xu Liu ◽  
Xiaoqiang Di ◽  
Jinqing Li ◽  
Huamin Yang ◽  
Ligang Cong ◽  
...  

Background: User behavior models have been widely used to simulate attack behaviors in the security domain. We revised all patents related to response to attack behavior models. How to decide the protected target against multiple models of attack behaviors is studied. Methods: We utilize one perfect rational and three bounded rational behavior models to simulate attack behaviors in cloud computing, and then investigate cloud provider’s response based on Stackelberg game. The cloud provider plays the role of defender and it is assumed to be intelligent enough to predict the attack behavior model. Based on the prediction accuracy, two schemes are built in two situations. Results: If the defender can predict the attack behavior model accurately, a single-objective game model is built to find the optimal protection strategy; otherwise, a multi-objective game model is built to find the optimal protection strategy. Conclusion: The numerical results prove that the game theoretic model performs better in the corresponding situation.


Author(s):  
Stuart Glennan

This chapter explores how mechanisms and their constituents can be classified into kinds. It argues for a weakly realist account of natural kinds—one which suggests that classification into kinds is based upon real similarities between instances of those kinds, but which denies that kinds have essences or have some reality apart from their instances. I introduce a models-first account of kinds, which suggests that two things are of the same kind to the extent that they can be represented by the same model. Because target entities can be represented by multiple models, they will belong to multiple kinds. I illustrate the approach by showing how the entities and activities that make up mechanisms can be classified into kinds.


Sign in / Sign up

Export Citation Format

Share Document