使用人工神经网络预测果果园中果Anastrepha fraterculus种群的季节性动态
Emílio de Souza Pimentel1, Jhersyka da Silva Paes2, Yuri Jivago Ramos3
1Dept of Entomology, Federal Univ of Viçosa, Viçosa, MG, Brazil. emilio.pimentelufv@gmail.com.
Neotropical entomology
|September 10, 2025
概括
人工神经网络 (ANN) 可以预测果园害虫Anastrepha fraterculus的季节动态. 该模型使用35天后的气象数据,准确预测害虫密度.
科学领域:
- 农业昆虫学 农业昆虫学
- 计算生物学 计算生物学
- 虫害管理 虫害管理 虫害管理
背景情况:
- 果 (Anastrepha fraterculus) 是商业果园的重要害虫.
- 预测季节性害虫种群动态对于有效的管理策略至关重要.
- 人工神经网络 (ANN) 为模拟复杂的生态现象提供了强大的工具.
研究的目的:
- 开发和验证一个人工神经网络 (ANN) 模型,用于预测果园中的Anastrepha fraterculus季节性种群动态.
- 确定影响害虫密度的关键气象因素.
- 评估模型在不同环境条件下的预测准确度.
主要方法:
- 从两个商业果园收集了八年来昆虫密度和气象数据.
- 开发和评估了1452个ANN,根据预测准确性和错误指标选择了最佳模型.
- 利用35天的延迟将气象数据输入到所选的ANN.
主要成果:
- 选择的ANN,结合35天后的气象数据,在预测和观察到的害虫密度之间实现了高相关性 (0.693).
- 该模型显示了低平方根平均验证误差 (0.066) 并在其隐藏层中显示了五个神经元.
- 风速和相对湿度与害虫密度正相关,而降水显示出负相关.
结论:
- 开发的ANN模型有效地预测了果园中的Anastrepha fraterculus密度.
- 该模型的准确性在不同的果园,作物阶段和年份中得到了验证,这表明其强度.
- 这种基于ANN的方法显示出在果种植中积极管理害虫的重大前景.
相关概念视频
00:53Migration
Migration is long-range, seasonal movement from one region or habitat to another. This common strategy, carried out by many different organisms around the world, is an adaptive response that typically corresponds to changes in an organism’s environment, like resource availability or climate. Migrations can involve huge groups of thousands of animals as well as single individuals traveling alone and can range from thousands of kilometers to just a few hundred meters.
01:25Arboviral Encephalitis
Arboviral encephalitis refers to brain inflammation caused by arthropod-borne viruses, particularly those transmitted through mosquito vectors. Among these, West Nile virus (WNV), a member of the Flaviviridae family, is a significant public health concern. WNV is an enveloped, positive-sense, single-stranded RNA virus. Human infection typically begins when an infected mosquito introduces the virus into the dermis during feeding. The primary transmission cycle involves birds as amplifying hosts...