Analytically determining frequency and amplitude of spontaneous alpha oscillation in Jansen's neural mass model using the describing function method

Spontaneous alpha oscillations are a ubiquitous phenomenon in the brain and play a key role in neural information processing and various cognitive functions.Jansen＇s neural mass model（NMM） was initially proposed to study the origin of alpha oscillations.Most of previous studies of the spontaneous alpha oscillations in the NMM were conducted using numerical methods.In this study,we aim to propose an analytical approach using the describing function method to elucidate the spontaneous alpha oscillation mechanism in the NMM.First,the sigmoid nonlinear function in the NMM is approximated by its describing function,allowing us to reformulate the NMM and derive its standard form composed of one nonlinear part and one linear part.Second,by conducting a theoretical analysis,we can assess whether or not the spontaneous alpha oscillation would occur in the NMM and,furthermore,accurately determine its amplitude and frequency.The results reveal analytically that the interaction between linearity and nonlinearity of the NMM plays a key role in generating the spontaneous alpha oscillations.Furthermore,strong nonlinearity and large linear strength are required to generate the spontaneous alpha oscillations.

A novel observer design method for neural mass models

Neural mass models can simulate the generation of electroencephalography（EEG） signals with different rhythms,and therefore the observation of the states of these models plays a significant role in brain research. The structure of neural mass models is special in that they can be expressed as Lurie systems. The developed techniques in Lurie system theory are applicable to these models. We here provide a new observer design method for neural mass models by transforming these models and the corresponding error systems into nonlinear systems with Lurie form. The purpose is to establish appropriate conditions which ensure the convergence of the estimation error. The effectiveness of the proposed method is illustrated by numerical simulations.

Analysis of the Mass Appraisal Model by Using Artificial Neural Network in Kaohsiung City

An accurate assessment of the property value is very important to make a deal, property tax, and mortgage for loan. The mass appraisal system has been developed in some foreign countries, especially in American for a long time. In Taiwan, we still have few experiences in using computer-assisted mass appraisal system, especially using artificial neural network （ANN）. This article has two objectives： （1） to illustrate application of ANN to the Kaohsiung property market by the method of back-propagation. The study is based on the properties data of sales price, we also use multiple regressions in the same data; （2） to evaluate the performance of two models by using the mean absolute percentage error （MAPE） and hit ratio （HR）. This paper finds that using artificial neural network （ANN） is able to overcome multiple regressions＇ methodological problems and also get better performance than multiple regression model （MRA）. These results are useful in helping local government to assess their assessment value.

Application of artificial neural networks in optimal tuning of tuned mass dampers implemented in high-rise buildings subjected to wind load

High-rise buildings are usually considered as flexible structures with low inherent damping. Therefore, these kinds of buildings are susceptible to wind-induced vibration. Tuned Mass Damper（TMD） can be used as an effective device to mitigate excessive vibrations. In this study, Artificial Neural Networks is used to find optimal mechanical properties of TMD for high-rise buildings subjected to wind load. The patterns obtained from structural analysis of different multi degree of freedom（MDF） systems are used for training neural networks. In order to obtain these patterns, structural models of some systems with 10 to 80 degrees-of-freedoms are built in MATLAB/SIMULINK program. Finally, the optimal properties of TMD are determined based on the objective of maximum displacement response reduction. The Auto-Regressive model is used to simulate the wind load. In this way, the uncertainties related to wind loading can be taken into account in neural network’s outputs. After training the neural network, it becomes possible to set the frequency and TMD mass ratio as inputs and get the optimal TMD frequency and damping ratio as outputs. As a case study, a benchmark 76-story office building is considered and the presented procedure is used to obtain optimal characteristics of the TMD for the building.

黄金矿山岩体质量分级知识库与PLS简化预测模型

针对黄金矿山工程岩体特征,分析了岩石单轴抗压强度、RQD值、节理结构面状态、节理结构面间距、地下水状态、节理结构面方向对工程影响和地应力值这7个主要因素对岩体稳定性的影响,对7个指标进行修正,建立了地下矿山M-RMR岩体质量评价指标体系。采用M-RMR岩体质量评价指标体系划分了焦家金矿直属矿区、寺庄矿区和望儿山矿区工程岩体质量等级,建立了焦家金矿地下矿山岩体质量与其影响因素的神经网络知识库模型,达到了焦家金矿工程岩体质量智能分级的目的。为简化M-RMR指标体系中指标数量,更利于实际应用,采用变量投影重要性指标VIP对7个指标所携带信息量的大小进行排序,并逐个删除不重要的指标,利用单因变量的偏最小二乘回归方法(PLS1)建立了精简指标的简化预测模型。简化预测模型可使用较少的评价指标对岩体质量给出准确的分级,具有实际使用价值。

基于AGA-RBF神经网络模型的叶绿素a质量浓度预测研究

叶绿素a质量浓度是预测湖泊水华形成的重要影响因子,但常用的径向基(radial basis function,RBF)神经网络存在容易陷入局部极值,导致预测精度欠佳.针对这一问题,采用自适应遗传算法(adaptive genetic algorithm,AGA)对RBF神经网络进行优化,构建基于AGA-RBF神经网络预测模型,以莆田东圳水库为应用案例,对叶绿素a质量浓度进行预测,通过采集到的数据对预测模型进行仿真,对比均方根误差(RMSE)、相对误差(RE)以及平均相对误差(MRE),验证改进后的AGA-RBF模型具有更好的预测精度,以期对叶绿素a质量浓度进行长期预测.

融合事理知识的群体性事件演化预测

[研究目的]当前,由社会矛盾和冲突所引发的网络群体性事件对公共安全构成了严重威胁,预测群体性事件的演化趋势对提升社会危机的防范与治理能力至关重要。[研究方法]提出了一种融合事理知识的群体性事件演化预测方法。该方法首先利用大语言模型蕴含的丰富事件知识及模型强大的生成能力来构建初始的事理图谱,通过结合真实新闻报道来验证图谱内容;其次,将事理图谱与图神经网络模型相结合,以获得更充分的事件语义表示;最后,基于上下文事件与候选事件间的语义相似度来预测未来可能发生的事件。[研究结论]研究表明,所提方法在事件演化预测的准确性和可解释性方面均显著优于参与比较的其他方法,验证了事理知识融合对揭示群体性事件演化模式的可行性和有效性。

重载列车运行过程的建模与RBFNN滑模控制

【目的】为解决重载列车在复杂线路条件下难以实现高精度轨迹跟踪控制的问题,提出了一种重载列车多质点模型和径向基函数神经网络滑模控制(RBFNNSMC)方法。【方法】首先,考虑空气制动和钩缓装置约束,建立重载列车多质点模型,并对人为测量误差和车辆参数差异等导致的模型不确定性问题,利用RBFNN对其进行估计。其次,设计一种非线性干扰观测器(NDO)对列车运行中受强风、雨雪等外界快时变干扰进行实时估计。然后,设计Lyapunov函数对整个系统进行稳定性证明。【结果】基于大秦线的实际线路数据,进行RBFNNSMC方法、PID方法和SMC方法的速度跟踪对比实验。仿真结果表明,RBFNNSMC方法的速度误差在±0.15 km/h以内,优于其他两种方法。加入NDO后,RBFNNSMC方法的抗干扰能力也更强。【结论】基于NDO的RBFNNSMC方法的跟踪精度相较于SMC方法在无干扰和受干扰情况下分别提升27.3%和28.9%,鲁棒性有所提升。

融合人脸图像深度和外观特征的BMI估计方法

身体质量指数(BMI)是人类健康重要指标。从2D正脸图像中估计3D人脸信息并提出一个端到端BMI估计框架,以进一步提高BMI估计性能。首先,计算人脸468个3D关键点,并根据关键点相对头部质心的深度绘制深度人脸图;其次,提取人脸图像的方向梯度直方图(HOG)并可视化以表示外观特征;最后,利用卷积神经网络(CNN)VGGNet、ResNet分别对深度人脸图和HOG进行特征提取,并使用Hadamard积融合2个骨干网络的特征以估计BMI。与目前已有方法的对比实验中,本文提出方法在2个公开数据集上的整体平均绝对误差(MAE)分别比最优结果低0.38和1。上述实验结果证明了本文提出的融合3D人脸图像深度和外观特征的BMI估计方法的有效性。

A generic method for rock mass classification

Rock mass classification(RMC) is of critical importance in support design and applications to mining,tunneling and other underground excavations. Although a number of techniques are available, there exists an uncertainty in application to complex underground works. In the present work, a generic rock mass rating(GRMR) system is developed. The proposed GRMR system refers to as most commonly used techniques, and two rock load equations are suggested in terms of GRMR, which are based on the fact that whether all the rock parameters considered by the system have an influence or only few of them are influencing. The GRMR method has been validated with the data obtained from three underground coal mines in India. Then, a semi-empirical model is developed for the GRMR method using artificial neural network(ANN), and it is validated by a comparative analysis of ANN model results with that by analytical GRMR method.

神经网络方法在核质量中的应用

基于两个具有代表性的宏观微观模型——有限程液滴模型(FRDM)和Weizs?cker-Skyrme模型(WS4),本文利用人工神经网络方法对模型所给出的数据进行了优化。加入神经网络方法后,FRDM所给出的结合能数据与2095个实验数据之间的均方根偏差从0.579 MeV降到0.354 MeV,WS4所给出的结合能数据与2095个实验数据之间的均方根偏差从0.292 MeV降到0.210 MeV。本文基于优化后的数据计算了Z=82同位素链的单中子分离能,FRDM和WS4单中子分离能的均方根偏差分别为39.9 keV和40.8 keV。此外,本文结合原始模型所给出的结合能数据,利用神经网络方法将数据进行了外推,FRDM和WS4在超重核区单中子分离能的均方根偏差分别为40.1 keV和188.1 keV。本文预测了新元素Z=119和Z=120同位素链的中子分离能,结果可为新元素合成的理论研究提供数据参考。

An ANN-Based Short-Term Temperature Forecast Model for Mass Concrete Cooling Control

Concrete temperature control during dam construction(e.g.,concrete placement and curing)is important for cracking prevention.In this study,a short-term temperature forecast model for mass concrete cooling control is developed using artificial neural networks(ANN).The development workflow for the forecast model consists of data integration,data preprocessing,model construction,and model application.More than 80000 monitoring samples are collected by the developed intelligent cooling control system in the Baihetan Arch Dam,which is the largest hydropower project in the world under construction.Machine learning algorithms,including ANN,support vector machines,long short-term memory networks,and decision tree structures,are compared in temperature prediction,and the ANN is determined to be the best for the forecast model.Furthermore,an ANN framework with two hidden layers is determined to forecast concrete temperature at intervals of one day.The root mean square error of the forecast precision is 0.15∘C on average.The application on concrete blocks verifies that the developed ANN-based forecast model can be used for intelligent cooling control during mass concrete construction.

考虑岩体结构面组合的区域崩塌灾害易发性评价

崩塌灾害是山区城镇和交通沿线主要的地质灾害类型,岩体优势结构面与斜坡临空面的空间组合关系是崩塌灾害失稳破坏的主要控制因素之一。针对平面滑移式、楔形体滑移式和坠落式崩塌灾害,综合考虑岩体优势结构面与斜坡临空面的空间组合模式以及岩体结构面的发育密度,提出斜坡单元内崩塌灾害发生概率的计算方法,分析崩塌灾害的易发斜坡坡向,构建适用于区域崩塌灾害易发性分析的岩体结构面组合模式评价指标体系,并以重庆市万州区铁峰乡为研究区,综合考虑崩塌灾害的岩性组合、地形坡度、植被指数(NDVI)、岩体优势结构面与斜坡坡向组合、地表水汇流影响和切坡建筑加载6个评价指标,以斜坡单元作为崩塌灾害易发性评价单元,采用人工神经网络模型对研究区崩塌灾害易发性进行定量评价。评价结果表明:地形坡度和岩体优势结构面与斜坡坡向组合是影响研究区崩塌灾害易发性的关键因素,预测模型的总体正确率AUC值为87.1%,说明模型的预测精度较高。该评价结果对研究区崩塌灾害风险管控具有指导意义。

1000 MW超临界燃煤发电机组脱硝喷氨自动控制方法研究

针对传统脱硝喷氨控制效果不佳的问题,提出了1000 MW超临界燃煤发电机组脱硝喷氨自动控制方法。基于BP神经网络建立1000 MW超临界燃煤发电机组脱硝喷氨系统模型,基于门控循环单元神经网络模型完成出口NOx质量浓度测量,通过混合-径向基神经网络完成1000 MW超临界燃煤发电机组脱硝喷氨的自动控制。测试结果表明,所提方法的发电机组脱硝喷氨自动控制效果良好,能够有效降低氨气逃逸量。

基于神经网络及机理分析的气力输送粉料质量流量软测量

本文提出了神经网络与机理分析结合的软测量方法 ,用以实现对气力输送系统中粉料质量流量的在线测量。通过实验验证 ,这种混合软测量方法是有效的。同时 ,与机理模型以及与基于标准神经网络的软测量方法的比较研究表明 。

食管癌血清差异蛋白质组测定结果分析

目的用表面增强激光解析电离飞行时间质谱技术(SELDI-TOF-MS)寻找食管癌(EC)患者血清中差异蛋白质组,并建立诊断模型,为其广泛用于临床诊断积累数据资料。方法用SELDI芯片检测食管癌及其相关人群的血清,Biomarker Wiz-ard Software软件筛选出差异蛋白,选择具有显著差异的蛋白质组建立人工神经网络诊断模型,经盲法验证后使用SPSS软件分析各个差异蛋白以及联合人工神经网络技术后的诊断效率,并鉴定相关蛋白。结果筛选出显著差异(P<0.01)的蛋白质组如下:在食管癌患者血清中表达增高蛋白组[(5 017.6±4.89)、(7 458.5±6.49)、(7 908.1±7.80)、(8 111±8.45)、(8 577±7.80)kD],表达降低蛋白组[(7 748.3±9.10)、(5 890.9±7.32)、(4 213.8±5.93)kD]。差异蛋白质组建立食管癌人工神经网络筛查模型和诊断模型,其灵敏度和特异性超过90%,盲法验证的效果好;人工神经网络预测输出值构成的ROC曲线下面积分别为81.6%和82.1%,经查阅数据库鉴定发现两种蛋白质。结论血清蛋白质谱指纹图结合人工神经网络技术进行蛋白质组学数据挖掘所建模型诊断价值较大,与临床诊断符合程度良好。

改进PSO算法结合FLANN在传感器动态建模中的应用

将改进的粒子群优化(PSO)算法和函数联接型神经网络(FLANN)相结合,实现传感器的动态线性建模。利用传感器的动态标定实验数据,首先训练FLANN神经网络,网络训练结束后的权值作为粒子群中某个粒子的初始值,而后利用改进的PSO算法继续寻优,得到的全局最优值即为所求的传感器动态模型的系数。实验结果表明,该方法结合了PSO和FLANN两者的优点,建模精度高。

基于神经网络的超声波轻质燃油质量流量测量方法研究

在对不同油品进行实验的基础上,分析了轻质燃油超声波传播速度、温度和密度之间的关系,建立了不同油品密度与温度和超声波传播速度之间人工神经网络模型。利用该模型,通过超声波速和温度测量,实现超声波流量计的直接质量计量,而不考虑油品批次和标准密度不同造成的影响。其密度测量误差小于0.24%,原理样机质量流量测量误差优于0.35%。

BP人工神经网络在岩体质量分级中的应用

通过金沙江某水电站的野外地质调查获取了丰富的基础地质资料后,针对其地质特点,提炼出岩石单轴抗压强度(R)、岩石质量指标(RQD)、岩体风化程度、节理组数(Jn)、节理粗糙系数(Jr)、节理蚀变系数(Ja)和地下水状态7个对岩体质量起控制作用的因素作为神经网络模型的输入变量。基于MATLAB软件平台设计的人工神经网络,通过具有较强代表性的116组样本数据的训练得到了稳定性好、可信度高的岩体质量分级网络模型。在对36组测试数据分析后,发现该模型的仿真结果具有较高的准确度和良好的吻合度,能够满足实际工程需要。

材料化学性质影响微生物污垢的神经网络分析

用神经网络技术对微生物垢诱导期和平均污垢量建立了评价及预测模型。对文献及作者实验获得的共 18个样本进行了拟合评价 ,结果表明 ,对诱导期 18个样本的评价结果与实测结果全部相符合 ,准确率达到 10 0 % ;对平均污垢量 ,有 15个样本的评价结果与实测结果相符合 ,准确率达到 83.3%。对 5个样本进行了预测 ,结果表明 ,对诱导期和平均污垢量 ,均有 4个样本的预测结果与实测结果相符 ,准确率达到 80 %。