人工智能培训

搜索

机器学习论文:基于大规模特征工程和国际流程反卷积的流感模型(Influenza Modeling Based on Massive Feature Engine

[复制链接]
77677 发表于 2019-12-9 13:04:26 | 显示全部楼层 |阅读模式
77677 2019-12-9 13:04:26 191 0 显示全部楼层
机器学习论文:基于大规模特征工程和国际流程反卷积的流感模型(Influenza Modeling Based on Massive Feature Engineering and  International Flow Deconvolution)在本文中,我们重点分析驱动流感蔓延的潜在因素以及减轻该疾病不利影响的可能政策。确切地说,我们首先调用离散傅里叶变换(DFT)得出流感活动的年度周期性区域结构,从而安全地将自己限制在年度流感行为的分析范围内。然后,我们从外部数据中收集了大量可能的区域性指标,这些指标可能与流感死亡率有关,例如消费,免疫,卫生,水质和其他指标,总大小为$ 1170。我们使用数据分析技术的组合从高维指标中提取重要特征,包括矩阵完成,支持向量机(SVM),自动编码器和主成分分析(PCA)。此外,我们将国际移民和贸易流建模为对区域流感活动的卷积,并解决了反卷积问题,即对线性回归的高阶扰动,从而将与流感死亡率相关的区域和国际因素分开。最后,原始模型和扰动模型都在区域示例中进行了测试,以验证我们的模型。与该策略有关,我们基于连接性数据以及先前提取的重要功能提出了一项建议,以减轻流感的影响,并有效地传播和执行该策略。我们得出结论,环境特征和经济特征对流感死亡率具有重要意义。该模型可以轻松地调整为其他类型的传染病模型。
In this article, we focus on the analysis of the potential factors drivingthe spread of influenza, and possible policies to mitigate the adverse effectsof the disease.To be precise, we first invoke discrete Fourier transform (DFT)to conclude a yearly periodic regional structure in the influenza activity,thus safely restricting ourselves to the analysis of the yearly influenzabehavior.Then we collect a massive number of possible region-wise indicatorscontributing to the influenza mortality, such as consumption, immunization,sanitation, water quality, and other indicators from external data, with $1170$dimensions in total.We extract significant features from the high dimensionalindicators using a combination of data analysis techniques, including matrixcompletion, support vector machines (SVM), autoencoders, and principalcomponent analysis (PCA).Furthermore, we model the international flow ofmigration and trade as a convolution on regional influenza activity, and solvethe deconvolution problem as higher-order perturbations to the linearregression, thus separating regional and international factors related to theinfluenza mortality.Finally, both the original model and the perturbed modelare tested on regional examples, as validations of our models.Pertaining tothe policy, we make a proposal based on the connectivity data along with thepreviously extracted significant features to alleviate the impact of influenza,as well as efficiently propagate and carry out the policies.We conclude thatenvironmental features and economic features are of significance to theinfluenza mortality.The model can be easily adapted to model other types ofinfectious diseases.机器学习论文:基于大规模特征工程和国际流程反卷积的流感模型(Influenza Modeling Based on Massive Feature Engineering and  International Flow Deconvolution)
URL地址:https://arxiv.org/abs/1912.02989     ----pdf下载地址:https://arxiv.org/pdf/1912.02989    ----机器学习论文:基于大规模特征工程和国际流程反卷积的流感模型(Influenza Modeling Based on Massive Feature Engineering and  International Flow Deconvolution)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

77677当前离线
新手上路

查看:191 | 回复:0

快速回复 返回顶部 返回列表