Aedes-AI: Neural network models of mosquito abundance.
Aedes-AI: Neural network models of mosquito abundance.
复制标题
DOI：
10.1371/journal.pcbi.1009467
复制
发表时间：
2021-11
影响因子：
4.3
通讯作者：
Lega J
中科院分区：
文献类型：
--
作者：
Kinney AC;Current S;Lega J
我们提出人工神经网络作为一个可行的替代蚊子丰度的机制模型。我们开发了前馈神经网络、长短期记忆递归神经网络和门控递归单元网络。我们评估了网络复制机制模型预测的蚊子种群时空特征的能力，并讨论了用强调特定动态行为的时间序列增强训练数据如何影响模型性能。最后，我们展望了这种无方程模型如何促进病媒控制或在任意空间尺度上估计疾病风险。埃及伊蚊每年通过基孔肯雅热、登革热和寨卡病毒等传染病影响数百万人。由于当地病媒水平需要足够高，才会发生相关的疫情，因此估计蚊子数量的能力是评估疾病风险的一个核心组成部分。蚊子景观模型（mools）是一种根据当地天气时间序列估计埃及伊蚊丰度的机制模型，能够重现监测数据中观察到的趋势。然而，将其扩展到大量位置是资源密集型的，需要高性能计算系统。在本文中，我们开发的人工神经网络模型比MoLS快得多，并且可以直接从当地天气数据产生丰度估计。这种方法减少了与估计当地蚊子水平相关的计算时间，从而允许这些预测的时空分辨率相应增加。我们比较了网络设计选择，包括架构和训练数据，以准确地再现MoLS估计和分析美国连续地点的模型性能。
We present artificial neural networks as a feasible replacement for a mechanistic model of mosquito abundance. We develop a feed-forward neural network, a long short-term memory recurrent neural network, and a gated recurrent unit network. We evaluate the networks in their ability to replicate the spatiotemporal features of mosquito populations predicted by the mechanistic model, and discuss how augmenting the training data with time series that emphasize specific dynamical behaviors affects model performance. We conclude with an outlook on how such equation-free models may facilitate vector control or the estimation of disease risk at arbitrary spatial scales. Aedes aegypti mosquitoes affect millions of people each year through infectious diseases such as chikungunya, dengue, and Zika. Because local vector levels need to be sufficiently high for associated outbreaks to occur, the ability to estimate mosquito abundance is a central component of assessing disease risk. The mosquito landscape model (MoLS) is a mechanistic model that estimates Aedes aegypti abundance from local weather time series, and is able to reproduce trends observed in surveillance data. However, scaling this up to a large number of locations is resource intensive, requiring a high-performance computing system. In this article, we develop artificial neural network models that are significantly faster than MoLS and can produce abundance estimates directly from local weather data. This approach reduces the computational time associated with estimating local mosquito levels, thereby allowing for a corresponding increase in the spatiotemporal resolution of these predictions. We compare network design choices, including architecture and training data, in their ability to accurately reproduce MoLS estimates and analyze model performance in locations across the contiguous United States.
登录
查看更多内容
人类的影响塑造了埃及埃及，登革热和黄热病蚊子的历史和最新进化。
DOI：
10.1111/evo.12281
发表时间：
2014-02
期刊：
Evolution; international journal of organic evolution
影响因子：
--
作者：
Brown JE;Evans BR;Zheng W;Obas V;Barrera-Martinez L;Egizi A;Zhao H;Caccone A;Powell JR
通讯作者：
Powell JR
DOI：
10.1002/joc.2312
发表时间：
2012-04-01
期刊：
INTERNATIONAL JOURNAL OF CLIMATOLOGY
影响因子：
--
作者：
Abatzoglou, John T.;Brown, Timothy J.
通讯作者：
Brown, Timothy J.
影响因子：
8.4
作者：
Ong, Adrian;Sandar, Mya;Sin, Leo Yee
通讯作者：
Sin, Leo Yee
影响因子：
3.2
作者：
Morin, Cory W.;Comrie, Andrew C.
通讯作者：
Comrie, Andrew C.
影响因子：
3.8
作者：
Hemme RR;Thomas CL;Chadee DD;Severson DW
通讯作者：
Severson DW