中国流行性腮腺炎发病率模型拟合及预测效果比较

背景　流行性腮腺炎（流腮）是中国极为严重的疾病。充分认识中国流腮的规律性并构建模型预测,对其预防和控制有重要意义。目的　评价季节性自回归移动平均模型（SARIMA）、指数平滑模型、SARIMA-广义回归神经网络（GRNN）组合模型和指数平滑-GRNN组合模型在流腮发病率拟合及预测中的应用效果。方法　利用全国2004年1月-2016年6月的流腮逐月发病率数据拟合、训练模型,建立SARIMA、指数平滑模型、SARIMA-GRNN组合模型和指数平滑-GRNN组合模型。预测2016年7-12月流腮的逐月发病率并与实际值比较,采用平均绝对误差百分比（MAPE）、平均误差率（MER）、均方误差（MSE）和平均绝对误差（MAE）评价模型拟合及预测效果。结果　SARIMA（0,0,2）（0,1,1）12为最优SARIMA模型;Holt-Winters相乘模型为最优指数平滑模型,SARIMA-GRNN组合模型和指数平滑-GRNN组合模型的SPREAD最优参数分别为0.026、0.031。SARIMA模型、指数平滑模型、SARIMA-GRNN组合模型和指数平滑-GRNN组合模型拟合的MAPE、MER、MSE和MAE依次分别为15.350%、14.976%、0.336、0.286,14.346%、14.249%、0.326、0.272,7.390%、6.320%、0.034、0.123,6.952%、5.776%、0.028、0.113。SARIMA模型、指数平滑模型、SARIMA-GRNN组合模型和指数平滑-GRNN组合模型预测的MAPE、MER、MSE和MAE依次分别为11.998%、12.260%、0.022、0.138,39.582%、38.462%、0.199、0.432,8.892%、9.677%、0.020、0.109,8.872%、9.672%、0.021、0.109。结论　指数平滑-GRNN组合模型为最优模型,拟合及预测效果最好,用于全国流腮发病率预测精度高;SARIMA-GRNN组合模型次之;SARIMA模型拟合及预测效果一般;指数平滑模型拟合效果较好,但预测效果较差。     

Performance evaluation of GRNN and ANFIS controlled DVR using machine learning in distribution network

A machine learning based generalized neural network estimator (GRNNE) and Takagi-Sugeno (T-S) fuzzy control system is implemented to accelerate the functional performance indices of dynamic voltage restorer (DVR). The GRNNE predictive model is recommended for the fast estimation of the reference load voltage under the distorted power supply. The fruit fly optimization learning strategy is employed to optimize the weights and smoothing parameters for the extraction of the reference voltage signals as well as unit vectors, resulting in the required sinusoidal load component. The dynamics in the DC-link voltage are optimized by the metaheuristic-based gray wolf optimization algorithm. The coefficients of the adaptive controller are updated automatically to achieve the best fitted adaptive neuro-fuzzy inference system predictive model with the least tracking voltage deviation error even in the presence of voltage disturbances. In comparison to classical techniques, the recommended neural-based approach offers a faster convergence speed with fewer parameters to tune, shorter training time, and lower risks of local entrapment. The performance metrics such as mean square error, root mean square error, mean absolute error, mean absolute percent error, coefficients of correlational and determination (R and R²) are used to evaluate the efficacy of the proposed controller and hence enhance the DVR performance. Finally, the simulation results of the hybrid approach confirm that the NN-based DVR estimator has proven its ability to alleviate voltage sensible issues at critical loads and outperform others in reducing power quality issues.     

Predicting Mechanical Properties of Cold-Rolled Steel Strips Using Micro-Magnetic NDT Technologies

Multiple micro-magnetic non-destructive testing (NDT) technologies are suitable candidates for predicting the mechanical properties of cold-rolled steel strips. In this work, based on magnetic domain dynamics behavior and magnetization theory, the correlation between electromagnetic characteristics extracted by multiple micro-magnetic NDT technologies and the influence factors was investigated. It was found that temperature and tension can subsequently affect the electromagnetic parameters by altering the domain structure and domain walls’ motion properties. Pearson’s correlation coefficients were employed to reflect the dependence of micromagnetic characteristics on influencing factors. The lift-off was determined as the largest influence factor among influence factors. A pseudo-static detection was reached by polynomial fitting, which could eliminate the influence of lift-off on the detection results. The number of training models was optimized, and the detection accuracy was improved via the improved Generalized Regression Neural Network (GRNN) model, based on the Gaussian Mixture Clustering (GMC) algorithm.     

ARIMA模型与GRNN模型对肺结核发病率预测的对比研究

目的 比较自回归移动平均(ARIMA)模型与广义回归神经网络(GRNN)模型对于肺结核发病率的预测性能.方法 根据我国2004年1月至2012年12月的肺结核逐月发病率数据资料,应用Eviews 7.0.0.1建立ARIMA模型,应用Matlab 7.1的神经网络工具箱建立GRNN模型;选取2013年肺结核逐月发病率数据对两种预测模型进行检验,比较预测结果.结果 ARIMA模型和GRNN模型的Theil不等系数(TIC)分别是0.034和0.059,说明ARIMA模型对我国2013年肺结核逐月发病率的拟合程度优于GRNN模型,ARIMA模型相对误差绝对值仅为GRNN模型的57.19％.结论 ARIMA预测模型更适合用于我国肺结核发病率的预测;建议尝试组合模型预测肺结核发病率.     

ARIMA模型与GRNN模型对性病发病率的预测研究

目的比较ARIMA模型与GRNN模型在性病发病率预测上的有效性。方法以新站区2005～2009年传染病疫情资料为基础,利用月发病率建立ARIMA(0,0,1)模型;将各月的发病率作为GRNN模型的输入,对应下个月发病率作为网络的输出,对样本进行训练,比较两模型的应用效果。结果合肥市新站区2005～2009年性病平均发病率达135.32/10万,并以年平均发展速度1.03缓慢上升。ARIMA(0,0,1)模型与GRNN模型的平均误差率分别是33.8%和27.0%;R2值分别是0.714和0.749。结论新站区性病呈平缓的波浪式上升,GRNN模型较适合该类疾病的发病率预测,建议公共卫生人员依据疫情预测及时做好防控工作。     

Parameters Identification of Rubber-like Hyperelastic Material Based on General Regression Neural Network

In this study, we present a systematic scheme to identify the material parameters in constitutive model of hyperelastic materials such as rubber. This approach is proposed based on the combined use of general regression neural network, experimental data and finite element analysis. In detail, the finite element analysis is carried out to provide the learning samples of GRNN model, while the results observed from the uniaxial tensile test is set as the target value of GRNN model. A problem involving parameters identification of silicone rubber material is described for validation. The results show that the proposed GRNN-based approach has the characteristics of high universality and good precision, and can be extended to parameters identification of complex rubber-like hyperelastic material constitutive.     

中医药对小儿急性非细菌感染性腹泻病的疗效及肠道微生态的影响

观察中医药疗法对小儿急性非细菌感染性腹泻病的疗效及肠道微生态的影响。方法:采用前瞻性多中心随机对照临床研究方法,将359例急性非细菌性腹泻病患儿分为试验A组、试验B组、对照组,分别采用中药汤剂、中药汤剂+小儿推拿、蒙脱石散(思密达)进行治疗,疗程3天。在治疗前、后进行疗效评价及应用实时荧光定量PCR法对粪便标本中双歧杆菌、大肠杆菌、乳酸菌的数量进行检测。结果:试验A组、试验B组的疗效均优于对照组(P0.05),试验A组、试验B组组间无差异(P0.05)。肠道微生态方面,除乳酸菌数量治疗后三组间及试验B组与对照组组间差异存在统计学意义(P0.05)外,其余乳酸菌、双歧杆菌、大肠杆菌数量治疗前后组内及治疗后组间比较均无统计学意义(P0.05);乳酸菌/大肠杆菌比值、双歧杆菌/大肠杆菌比值治疗后组间比较均无统计学意义(P0.05)。结论:中医药疗法对小儿急性非细菌感染性腹泻病具有良好的治疗作用,其短期内对肠道双歧杆菌、大肠杆菌、乳酸菌数量影响不明显。     

基于奇异谱分析的ARIMA模型在山西省流感预测中的应用

目的 分析基于奇异谱分析（singular spectrum analysis, SSA）的自回归移动平均模型（Autoregressive integrated moving average, ARIMA）模型预测流感样病例 (influenza like illness, ILI) 发病趋势的可行性，为流感防控工作提供合理的预测方法。 方法 利用山西省2010年第14周-2017年第13周的流感监测资料以不同长度配比的训练集、测试集构建SSA-ARIMA模型，并与ARIMA、BP神经网络（Back propagation neural network, BPNN）、广义回归神经网络（General Regression Neural Network, GRNN）模型进行比较。采用平均绝对误差（Mean Absolute Error，MAE）、均方误差（Mean Squared Error，MSE）、均方根误差（Root Mean Squared Error，RMSE）比较各模型预测效果。 结果 模型拟合方面，SSA-ARIMA模型在预测未来一个月发病趋势时的MAE、MSE、RMSE分别为0.163、0.061、0.248；预测六个月时分别为0.161、0.061、0.248；预测一年时分别为0.168、0.066、0.256；均低于ARIMA、BPNN、GRNN。模型预测方面，在预测未来一个月发病趋势时的MAE、MSE、RMSE分别为0.056、0.005、0.068；预测六个月时分别为0.189、0.081、0.285；预测一年时分别为0.210、0.075、0.273；也均低于ARIMA、BPNN、GRNN。 结论 SSA-ARIMA模型对山西省ILI的预测效果优于ARIMA、BPNN、GRNN，可为流感预测提供科学依据。     

基于GRNN的组合预测模型在传染病发病率预测中的应用

目的研究基于GRNN的组合预测模型拟合传染病发病率的优越性和不足。方法以浙中某市1998—2008年的肺结核发病率为研究资料,分别构建了灰色模型和ARIMA模型,以这两种模型为基础构建了基于GRNN的组合预测模型。结果残差修正GM(1,1)模型、ARIMA(1,0,1)*(1,1,0)12模型、基于GRNN的组合预测模型的MSE,MAE,MAPE和MER分别为37.451,5.692,53.69%,48.51%;18.509,3.761,35.13%,32.05%;9.961,2.571,25.6%,21.9%。结论基于GRNN的组合预测模型的预测精度优于两种单项模型。     

COVID-19 mortality rate prediction for India using statistical neural networks and gaussian process regression model

The primary purpose of this research is to identify the best COVID-19 mortality model for India using regression models and is to estimate the future COVID-19 mortality rate for India. Specifically, Statistical Neural Networks (Radial Basis Function Neural Network (RBFNN), Generalized Regression Neural Network (GRNN)), and Gaussian Process Regression (GPR) are applied to develop the COVID-19 Mortality Rate Prediction (MRP) model for India. For that purpose, there are two types of dataset used in this study: One is COVID-19 Death cases, a Time Series Data and the other is COVID-19 Confirmed Case and Death Cases where Death case is dependent variable and the Confirmed case is an independent variable. Hyperparameter optimization or tuning is used in these regression models, which is the process of identifying a set of optimal hyperparameters for any learning process with minimal error. Here, sigma (σ) is a hyperparameter whose value is used to constrain the learning process of the above models with minimum Root Mean Squared Error (RMSE). The performance of the models is evaluated using the RMSE and ''R2 values, which shows that the GRP model performs better than the GRNN and RBFNN.