预测误差

RBF网络的预测误差小于8%。

The prediction error of RBF network was less than 8 % .

作为多元统计过程控制方法中的常用统计量，平方预测误差（SPE）的变化规律有待深入研究。

Squared prediction error ( SPE ) statistic is frequently used in multivariate statistical process control and its law of variation needs further investigating .

然后通过预测误差的Context模型进一步去除相邻像素之间的相关性；

After that , the context model is applied to further remove the correlation between neighbor pixels .

该学习算法的收敛速度比BP学习算法快，且比并行递推预测误差学习算法有更好的数值稳定性。

It has better convergence than the back propagation learning algorithm and better stability than the parallel recursive prediction error learning algorithm .

用Bootstrap法估计均方预测误差

The Estimation of mean Square Forecast Error by Bootstrap and Method

基于LPC预测误差DWT的语音基音检测

Pitch Detection of Speech Signals Based on LPC Predictive Error DWT

基于MCP预测误差VaR值的日前现货市场竞价模型研究

Research on Day-ahead Power Market Bidding Model Based on VaR of MCP Forecasted Error

补偿器中网络延时预测误差是导致系统不稳定的主要原因，采用RBF神经网络对延时进行预测，该方法提高了延时预测精度，改善了系统的整体性能。

In the compensator predication errors of internet-induced time delays are the main cause to induce the system instability .

同时将BP网络累积预测误差标准差作为自适应遗传算法的适应度函数，并作为遗传终止的判定准则之一。

Meanwhile , AGA fitness function is defined as the standard deviation of BP network 's accumulation forecast errors and also one of genetic termination criteria .

经过与线性核函数及Sig-moid核函数的对比，选用基于径向基函数（RBF）作为核函数，在分析预测误差和模型参数关系的基础上，选择了合适的参数；

By analyzing the relationship between the error margin of prediction and the model parameters , the proper parameters were chosen .

R2（判定系数）接近于1，表明预测误差较小；

The determination coefficients was almost one , which showed the least prediction error .

研究方法：主成分分析法。利用主元相关变量残差统计量代替平方预测误差Q统计量，并采用累积方差贡献率及复相关系数确定PCA模型的主元数。

Method of principal components was employed . The principal components in PCA is decided by virtue of ( cumulative ) percent variance and multi-correlation coefficients .

给出了条件均方预测误差公式和Bootstrap法条件均方预测误差公式。

Conditional mean square forecast error formulas and bootstrap procedure dealing with conditional mean square forecast error formulas are given .

文章运用ARIMA过程的截尾形式为校正公式提供一处证明，并给出预测误差的方差。

A method is proposed to prove the updated forecast function in the truncated form for ARIMA processes .

本文提出了一种新的GM（1,1）模型，可以在灰度图像无损压缩中实现预测误差编码。

In this paper we propose a new GM ( 1,1 ) model , which can implement the predictive error coding in a lossless compression of gray scale images .

提出了一种新颖的具有预测误差校正的多变量广义预测控制（GPC）与动态矩阵控制（DMC）混合集中优化控制算法。

In this paper , a novel multivariable GPC and DMC mixed centralized optimization control algorithm with predictive error correcting is presented .

若将叶面吸收因子耦合到限制分配模型当中，则能显著降低茎叶中4种PAHs含量的预测误差。

If the impact of foliar uptake on model performance was taken into account , the differences for all the four PAHs would be greatly reduced .

本文用平方预测误差SPE这个统计量，来检测系统是否出现故障。

In this paper , squared prediction error ( SPE ), a statistic variable , is used to detect whether the system is faulty .

针对传统的主元分析（PCA）的T~2和平方预测误差（SPE）检验所提供的信息并不一致的缺陷，提出一种改进的PCA方法。

The information provided by T ~ 2 and squared prediction error ( SPE ) test of principal component analysis ( PCA ) is not cor - responding .

基于采集到的试验输入、输出数据对序列，采用基于LM准则函数寻优算法的预测误差法对各通道进行了模型参数辨识。

The model parameters are identified by using predicted-error method , which is based on the LM algorithm used for the criterion function optimization .

实例分析表明，所建模型预测误差均较小，好于门限自回归模型，BP神经网络模型和ELMAN神经网络模型。

The prediction precision of this RBF model is higher than that of the threshold auto-regression model , the BP artificial neural network model and the ELMAN artificial neural network model .

本文将推导一种基于对AR参数求向前向后预测误差最小二乘解的非递推建模算法（NLS）。

A new fast nonrecursive algorithm for AR modeling will be introduced , based on the least squares solution for AR parameters using forward and backward linear prediction .

计算不同预测误差百分比的符合率及其95%可信区间，判断预测的准确程度，验证PPK模型。

The coincidence rates and their 95 % confidence intervals of different percentages of prediction error were calculated .

广义预测误差模型及其GLMS自适应算法

Generalized prediction error model and its GLMs adaptive algorithm

Burg提出一种递推算法，不需要计算自相关，而是用使前向与后向预测误差能量之和最小的方法求出模型的参数。

Burg has advanced an algorithm , which computes the parameters by making the sum of forward and backward forecasting errors minimum .

将差值扩展技术应用于彩色图像，提出一种基于预测误差差值扩展和最低有效位（LSB）替换的彩色图像可逆数据隐藏算法。

This paper proposed a reversible data hiding algorithm for color images based on prediction error difference expansion and Least Significant Bit ( LSB ) replacement .

在供应充足的情况下，提出实施VMI计划抑制牛鞭效应，通过比较VMI实施前后供应链中的牛鞭效应以及供应商对需求的预测误差，发现VMI计划的实施对牛鞭效应的抑制确实比较有效。

It proves that VMI plan would also help restrain bullwhip effect . In supply deficiency condition , retailers intend to overestimate demands , which could account for bullwhip effect increase .

简要回顾了该领域国内外的研究成果，并分析总结了疲劳理论知识，针对传统疲劳寿命预测误差较大的问题，将隶属函数引入到寿命估算中，形成基于模糊修正Miner法则的疲劳寿命估算方法。

For the large error in traditional fatigue life prediction , membership function to the life estimation was introduced in order to form fatigue life estimation method based on fuzzy revised Miner rule .

本文采用权值可在线调整的动态补偿神经网络（动态BP网络）对模型预测误差进行拟合，从而显著提高了基于线性模型的非线性广义预测（GPC）的预测精度，增强了算法的鲁棒性。

A Generalized predictive control algorithm based on error correction using dynamically compensating neural network is proposed to approximate the modeling error . So it improves the predictive accuracy of GPC based on linear model and buildups systems robustness .

DPCM图象传输系统中，对量化后的预测误差所采取的代码分配方案影响到系统对信道误码的敏感程度。

In a DPCM image transmission system the codeword assigment for quantized predictive differences has influence on the performance of channel error sensitivity of the system .