共查询到20条相似文献，搜索用时 31 毫秒
1.
In order to investigate the eutrophication degree of Yuqiao Reservoir, a hybrid method, combining principal component regression (PCR) and artificial neural network (ANN), was adopted to predict chlorophyll-a concentration of Yuqiao Reservoir’s outflow. The data were obtained from two sampling sites, site 1 in the reservoir, and site 2 near the dam. Seven water variables, namely chlorophyll-a concentration of site 2 at time t and that of both sites 10 days before t, total phosphorus(TP), total nitrogen(TN), dissolved oxygen(DO), and temperature from January 2000 to September 2002, were utilized to develop models. To remove the collinearity between the variables, principal components extracted by principal component analysis were employed as predictors for models. The performance of models was assessed by the square of correlation coefficient, mean absolute error (MAE), root mean square error (RMSE) and average absolute relative error (AARE). Results show that the hybrid method has achieved more accurate prediction than PCR or ANN model. Finally, the three models were applied to predicting the chlorophyll-a concentration in 2003. The predictions of the hybrid method were found to be consistent with the observed values all year round, while the results of PCR and ANN models did not fit quite well from July to October. 相似文献
2.
Regression-Based Artificial Neural Network Methodology In Response Surface Methodology 总被引：3，自引：0，他引：3
Response surface methodology (RSM) is an important tool for process parameter optimization, robust design and other quality improvement efforts. When the relationship between influential input variables and output response is very complex, it‘ s hard to find the real response surface using RSM. In recent years artificial neural network (ANN) has been used in RSM. But the classical ANN does not work well under the constraints of real applications. An algorithm of regression-based ANN(R-ANN) is proposed in this paper, which is a supplement to the classical ANN methodology. It makes network closer to the response surface, so that training time is reduced and robustness is strengthened. The procedure of improving ANN by regressions is described and the comparisons among R-ANN, RSM and classical ANN are computed graphically in three examples. Our research shows that the R-ANN methodology is a good supplement to the RSM and classical ANN methodology, which can yield lower standard error of prediction under conditions that the scope of experiment is rigidly restricted. 相似文献
3.
Electrical discharge machining （EDM） process, at present is still an experience process, wherein selected parameters are often far from the optimum, and at the same time selecting optimization parameters is costly and time consuming. In this paper, artificial neural network （ANN） and genetic algorithm （GA） are used together to establish the parameter optimization model. An ANN model which adapts Levenberg-Marquardt algorithm has been set up to represent the relationship between material removal rate （MRR） and input parameters, and GA is used to optimize parameters, so that optimization results are obtained. The model is shown to be effective, and MRR is improved using optimized machining parameters. 相似文献
4.
In psychological, social, behavioral, and medical studies, hidden Markov models (HMMs) have been extensively applied to the simultaneous modeling of heterogeneous observation and hidden transition in the analysis of longitudinal data. However, the majority of the existing HMMs are developed in a parametric framework without latent variables. This study considers a novel semiparametric HMM, which comprises a semiparametric latent variable model to investigate the complex interrelationships among latent variables and a nonparametric transition model to examine the linear and nonlinear effects of potential predictors on hidden transition. The Bayesian P-splines approach and Markov chain Monte Carlo methods are developed to estimate the unknown, a Bayesian model comparison statistic, is employed to conduct model comparison. The empirical performance of the proposed methodology is evaluated through simulation studies. An application to a data set derived from the National Longitudinal Survey of Youth is presented. 相似文献
5.
刘谢进 《楚雄师范学院学报》2006,21(12):47-49
在给定的权回归模型下,讨论了最小二乘估计、最优加权最小二乘估计和线性无偏最小方差估计的性能比较,得出了在随机误差方差矩阵可逆条件下,可算出最优加权最小二乘估计与线性无偏最小方差估计误差方差阵的差表达式,并在一定条件下,两者趋于一致。 相似文献
6.
分别采用人工神经网络BP算法（网络结构为3-9-1）和线性回归分析方法对17个4-X-N-Y-6-氮杂雄-4-烯-3-酮衍生物在4℃时与小牛子宫雌激素受体的亲合力参数Iog1/K（iKi为衍生物对3BHSD的抑制常数）与分子的范德华体积V、最高被占据分子轨道能量EHOMO和9号碳原子的净电荷Q之间建立了QSAR模型,ANN模型的相关系数R=0.9999,标准偏差SD=0.0014。MLR模型的相关系数R=0.9470,标准偏差SD=0.4459。结果表明人工神经网络是一种比较精密的拟合方法,具有良好的预测效果。 相似文献
7.
讨论了输入、输出及回归系数都是LR-型模糊数的模糊线性回归模型参数估计的加权最小二乘法.该方法根据决策者对训练数据的置信度对观测数据设置不同的权重,从而得到能有效抵御异常值干扰的预测模型. 相似文献
8.
应用人工神经网络（Artificial Neural Network，ANN）算法对MIT-BIH心电数据库中的数据进行检测，对波形辨识算法做初步研究。设计三层神经网络结构：输入层、隐含层和输出层。从心电信号中提取4项特征参数作为输入层的输入量，并对MIT-BIH心电数据库中的15例心电数据进行了检测。表明该算法对QRS波总体检测灵敏度为98．96％，检测真阳性率为99．93％，对室性异位博动检测灵敏度为94．97％，检测真阳性率为98．72％。实验证实该神经几乎络算法对心电波形辨识非常有效。 相似文献
9.
WANG Rui-min CAO Guang-yi ZHU Xin-jian 《浙江大学学报(A卷英文版)》2007,8(5):741-747
Model and simulation are good tools for design optimization of fuel cell systems. This paper proposes a new hybrid model of proton exchange membrane fuel cell (PEMFC). The hybrid model includes physical component and black-box com-ponent. The physical component represents the well-known part of PEMFC, while artificial neural network (ANN) component estimates the poorly known part of PEMFC. The ANN model can compensate the performance of the physical model. This hybrid model is implemented on Matlab/Simulink software. The hybrid model shows better accuracy than that of the physical model and ANN model. Simulation results suggest that the hybrid model can be used as a suitable and accurate model for PEMFC. 相似文献
10.
提出一种同时考虑解释性和精确性的模糊建模方法. 首先分析影响模糊模型解释性的主要因素, 然后利用启发式搜索策略实现输入变量选择, 利用模糊聚类算法和最小二乘辨识模糊模型. 随后以输入变量数目和模糊规则数目的乘积衡量可解释性, 以均方误差衡量精确性, 并据此定义模型选择目标函数. 最后给定最大最小的输入变量数目和规则数目, 辨识得到一组模糊模型, 利用模型选择目标函数, 选择最优的模糊模型, 并采用遗传算法进行优化, 达到解释性与精确性的折衷. 煤气炉仿真例子验证了该方法的有效性. 相似文献
11.
提出了基于人工神经网络（ＡｒｔｉｆｉｃｉａｌＮｅｕｒａｌＮｅｔｗｏｒｋｓ）对动力结构进行系统辨识的方法，即应用人工神经网络预测结构地震响应．采用ＢＰ算法的前馈网络（简称ＢＰ网络）对剪切模型结构进行系统辨识．首先用实际地震波及相应的模拟地震响应训练本文提出的ＢＰ网络，然后用“已学会”的ＢＰ网络预测其它地震波激励下的结构地震响应．还讨论了网络拓扑结构、输入单元数等对网络学习和预测的影响．通过本文可以发现，合适的人工神经网络结构能准确地辨识结构动力特性和预测结构动力响应 相似文献
12.
蒋惠凤 《温州职业技术学院学报》2014,(1):57-60
为提高负荷预测精度,将主成分回归（PCR）、偏最小二来回归（PLSR）与反向传播神经网络（BPNN）相结合,分别建立基于PCR和PLSR及与神经网络耦合的年用电量预测模型.结果表明,以PCR和PLSR方法提取成分作为神经网络的输入,以实际用电量作为输出,建立的PC-BPNN和LV-BPNN非线性预测模型拟合优度优于PCR和PLSR线性预测模型.从检验四个预测模型的预测效果看,线性预测模型的预测值均高于实际值,非线性预测模型的预测值均低于实际值. 相似文献
13.
王玉梅 《通化师范学院学报》2011,32(8):7-8
对于自变量具有多重线性相关性的多元线性回归,文中分别用主成分回归和PLSI回归对同一适当的问题进行建模分析,得出PLSI回归优于主成分回归的结论． 相似文献
14.
从神经网络结构设计问题出发,提出一种确定神经网络最优隐节点个数的新方法.该算法首先按照等差数列增加隐节点,确定最优隐节点个数的范围;然后利用折半删减法确定最优隐节点个数.数值实验表明该算法在保持良好泛化能力的同时能自适应地、快速有效地确定网络最小隐神经元数目. 相似文献
15.
Karl-Heinz Leitner Julia Prikoszovits Michaela Schaffhauser-Linzatti Rainer Stowasser Karin Wagner 《Higher Education》2007,53(4):517-538
This paper explores the performance efficiency of natural and technical science departments at Austrian universities using
Data Envelopment Analysis (DEA). We present DEA as an alternative tool for benchmarking and ranking the assignment of decision-making
units (organisations and organisational units). The method applies a multiple input and output variables approach, which is
a clear advantage to other approaches using simple performance ratios. To deliver reasonable results, suitable input and output
variables have been determined in a previous step using correlation analyses and OLS regression. The results validate the
methods applied, and reveal performance differences and scale effects. The use of multiple output variables enables the revealing
of detailed improvement or reduction amounts of each input and output of the evaluated units and furthermore for identifying
the specialisation of teaching, research, and industrial cooperation. We find significant evidence that the size of a department
influences its overall and specialisation performance; both small and large departments perform above average, which proves
that simple linear scale effects do not exist. 相似文献
16.
In this study, an artificial neural network （ANN） model for studying the strength properties of steel fiber reinforced concrete （SFRC） containing fly ash was devised. The mixtures were prepared with 0 wt%, 15 wt%, and 30 wt% of fly ash, at 0 vol.%, 0.5 vol.%, 1.0 vol.% and 1.5 vol.% of fiber, respectively. After being cured under the standard conditions for 7, 28, 90 and 365 d, the specimens of each mixture were tested to determine the corresponding compressive and flexural strengths. The pa- rameters such as the amounts of cement, fly ash replacement, sand, gravel, steel fiber, and the age of samples were selected as input variables, while the compressive and flexural strengths of the concrete were chosen as the output variables. The back propagation learning algorithm with three different variants, namely the Levenberg-Marquardt （LM）, scaled conjugate gradient （SCG） and Fletcher-Powell conjugate gradient （CGF） algorithms were used in the network so that the best approach can be found. The results obtained from the model and the experiments were compared, and it was found that the suitable algorithm is the LM algorithm. Furthermore, the analysis of variance （ANOVA） method was used to determine how importantly the experimental parameters affect the strength of these mixtures. 相似文献
17.
Cheryl L. Somers Rachel L. Goutman Angelique Day Oliva Enright Shantel Crosby Heather Taussig 《Psychology in the schools》2020,57(12):1845-1863
Research shows that youth in foster care experience poor academic performance and disciplinary actions in school more frequently than do non-foster care youth. The purpose of this cross-sectional study was to further examine youth in foster care and the relationship between individual/intrapersonal factors (future orientation and school connectedness) and exosystem factors (number of placement and school moves) and academic performance (grades) and disciplinary referrals among 363 youth (9–11 years of age; males = 52.9%). Controlling for key variables, hierarchical linear regression analysis was utilized to understand how well students' school connectedness, future outlook, number of placement changes, and number of school moves predicted academic and disciplinary outcomes. Beyond the variance explained by control variables, school connectedness made a significant contribution to this model. Results are discussed in the context of implementing interventions that foster school connectedness among this vulnerable population. 相似文献
18.
In underdetermined blind source separation, more sources are to be estimated from less observed mixtures without knowing source signals and the mixing matrix. This paper presents a robust clustering algorithm for underdetermined blind separation of sparse sources with unknown number of sources in the presence of noise. It uses the robust competitive agglomeration (RCA) algorithm to estimate the source number and the mixing matrix, and the source signals then are recovered by using the interior point linear programming. Simulation results show good performance of the proposed algorithm for underdetermined blind sources separation (UBSS). 相似文献
19.
A new customization approach based on support vector regression (SVR) is proposed to obtain individual headrelated
impulse response (HRIR) without complex measurement and special equipment. Principal component analysis (PCA) is
first applied to obtain a few principal components and corresponding weight vectors correlated with individual anthropometric
parameters. Then the weight vectors act as output of the nonlinear regression model. Some measured anthropometric
parameters are selected as input of the model according to the correlation coefficients between the parameters and the weight
vectors. After the regression model is learned from the training data, the individual HRIR can be predicted based on the
measured anthropometric parameters. Compared with a back-propagation neural network (BPNN) for nonlinear regression,
better generalization and prediction performance for small training samples can be obtained using the proposed PCA-SVR
algorithm. 相似文献
impulse response (HRIR) without complex measurement and special equipment. Principal component analysis (PCA) is
first applied to obtain a few principal components and corresponding weight vectors correlated with individual anthropometric
parameters. Then the weight vectors act as output of the nonlinear regression model. Some measured anthropometric
parameters are selected as input of the model according to the correlation coefficients between the parameters and the weight
vectors. After the regression model is learned from the training data, the individual HRIR can be predicted based on the
measured anthropometric parameters. Compared with a back-propagation neural network (BPNN) for nonlinear regression,
better generalization and prediction performance for small training samples can be obtained using the proposed PCA-SVR
algorithm. 相似文献
20.
提出了一种优化的迭代降维算法求解混合交通网络设计问题. 混合(连续/离散) 交通网络设计问题常表示为一个带均衡约束的数学规划问题,上层通过新建路段和改善已有路段来优化网络性能,下层是一个传统的 Wardrop 用户均衡模型. 迭代降维算法的基本思想是降维,先保持一组变量(离散/连续) 不变,交替地对另一组变量(连续/离散) 实现最优化. 以迭代的形式反复求解连续网络设计和离散网络设计问题,直至最后收敛到最优解. 通过一个数值算例对算法的效果进行了验证. 相似文献