絕對值誤差 的英文怎麼說

中文拼音 [juéduìzhíwùchā]
絕對值誤差 英文
absolute error
  • 絕 : Ⅰ動詞（斷絕） cut off; sever Ⅱ形容詞1 （完全沒有了; 窮盡; 凈盡） exhausted; used up; finished 2 ...
  • 對 : Ⅰ動詞1 （回答） answer; reply 2 （對待; 對付） treat; cope with; counter 3 （朝; 向; 面對） be tr...
  • 誤 : Ⅰ名詞（錯誤） mistake; error Ⅱ動詞1 （弄錯） mistake; misunderstand 2 （耽誤） miss 3 （使受損害...
  • 差 : 差Ⅰ名詞1 （不相同; 不相合） difference; dissimilarity 2 （差錯） mistake 3 [數學] （差數） differ...
  • 絕對值 : [物理學] [數學] absolute value; absolute; modulus絕對值符號 [自動化] absolute value sign; 絕對值...
  • 絕對 : absolute
  • 誤差 : error
  1. Compared with the classical bp algorithm, robust adaptive bp algorithm possesses some advantages as following : ( 1 ) increasing the accuracy of the network training by means of using both the relative and absolute residual to adjust the weight values ; ( 2 ) improve the robustness and the network convergence rate through combining with the robust statistic technique by way of judging the values of the samples " relative residual to establish the energy function so that can suppress the effect on network training because of the samples with high noise disturbances ; ( 3 ) prevent entrapping into the local minima area and obtain the global optimal result owing to setting the learning rate to be the function of the errors and the error gradients when network is trained. the learning rate of the weights update change with the error values of the network adaptively so that can easily get rid of the disadvantage of the classical bp algorithm that is liable to entrap into the local minima areas

    與基本bp演算法相比，本文提出的魯棒自適應bp演算法具有以下優點： （ 1 ）與魯棒統計技術相結合，通過訓練樣本相對偏差的大小，確定不同訓練樣本對能量函數的貢獻，來抑制含高噪聲干擾樣本對網路訓練的不良影響，從而增強訓練的魯棒性，提高網路訓練的收斂速度； （ 2 ）採用相對偏差和絕對偏差兩種偏差形式對權值進行調整，提高了網路的訓練精度； （ 3 ）在採用梯度下降演算法對權值進行調整的基礎上，通過將學習速率設為訓練誤差及誤差梯度的特殊函數，使學習速率依賴于網路訓練時誤差瞬時的變化而自適應的改變，從而可以克服基本bp演算法容易陷入局部極小區域的弊端，使訓練過程能夠很快的「跳出」局部極小區域而達到全局最優。
  2. The difference of two methods is relatively small and data of velocity accord preferably, the least absolute error is about 0. 001m / s and the least relative error is 0. 49 % in no boundary position, absolute error is big in the boundary positions which arouse measurement error

    將數值模擬的結果和數字圖像處理的結果相比較分析，發現兩種方法之間的速度差別較小，在非邊界點處兩種方法的速度數據符合相對較好，最小絕對誤差為0 . oolm / s ，最小相對誤差為0 . 49 % 。
  3. During the course of the research, the criterions of the interpolation effect are mean error ( me ), mean absolute error ( mae ), root mean squared interpolation error ( rmse ) and the difference of mean square deviation between the measured and the estimated surface air temperature. the conclusions are as follows : ( 1 ) by contrasting the gaussian weighted model associated with the error modification with the gaussian weighted model, the error modification is proved to considerably ameliorate the precision of spatial interpolation ; ( 2 ) on the base of the gaussian weighted model, taking altitudinal effect into account can reflect the trend in which temperature changes according to the topographic altitude and may ameliorate the precision of spatial interpolation correspondingly and apparently, which indicates that topographical effect on the preciseness of spatial interpolation can not be disregarded in terms of the region with complicated topography ; ( 3 ) the map of daily surface air temperature distribution, using the modified gaussian weighted model a and b, can accurately reflect the temperature - changing - with - topographical - altitude trend. among them, the better is the model a, whose me is below 0. 03 ?

    在此過程中，採用平均誤差（ me ） ，平均絕對誤差（ mae ） ，插值平均誤差平方的平方根（ rootmeansquaredinterpolationerror ，簡稱rmsie ） ，插值前後測站要素值的均方差（ meansquaredeviation ，簡稱msd ）差值作為判定插值效果的標準，得出如下結論：通過高斯權重法與結合逐步訂正的高斯權重法的對比，說明結合逐步訂正方案的高斯權重法可大大提高地面日氣溫的插值精度；在高斯權重法中加入海拔影響項可以反映出溫度隨地形高度的變化趨勢，同時也能較大地提高地面日氣溫的空間插值精度，說明在地形復雜的區域，地形影響在插值精度中是不可忽略的；對于高斯權重法的兩種改進方案得到的地面日氣溫分布圖都能很好地反映出表面大氣氣溫隨地形高度的變化趨勢。
  4. It has been proved in many examples that the bp algorithm based in the square sum of the relative error is better than the conventional bp method

    通過大量算例檢驗證實，在基於相對誤差平方和為檢驗標準前提下，利用所給演算法求得的擬合值或預報結果優于傳統的基於絕對誤差平方和作為目標函數的bp演算法所得結果。
  5. The probability of normal r. v. with absolute value less then and equal to one, two, and three unit standard deviations are calculated from error function, the cumulate probability distribution of normal density

    常態分佈隨機變數絕對值小於等於一個、兩個、三個標準差之機率分別代入累積機率分佈函數（誤差函數）算出其對應之機率。
  6. Similar to the famous von - mise condition on the extreme value theory of 1 - max style, the judge condition that absolutely continuous distribution function is in the domain of attraction of p - max style distribution function is given. at the same time, the error inequality between samples and true values is obtained, and almost sure convergence theorems on the extreme value theory of p - max style are also given

    類似於l - max型極值理論中著名的von - mise條件，本文給出了絕對連續分佈函數f落在p - max型極值分佈函數的吸收域中的判斷條件，給出了樣本與真值的誤差不等式，並給出了關于p - max型極值理論的幾乎處處收斂定理。
  7. Numerical examples are given to illustrate that the norm errors of the approximate algorithm is generally small, while the smaller the absolute value is, the greater the component errors will be

    數值算例表明：該近似演算法的范數誤差一般較小，而絕對量較小的分量誤差較大。
  8. From theoretical analysis, we know the existing demodulation methods have limitations as following : one is that the subtraction of the two signals frequencies will display as the result of demodulation when we demodulate two time - domain adding signals without modulating information ( fault information ) ; the other one is that aliasing phenomenon will occur as a result of getting absolute value, detection or square in the process of generalized demodulation analysis, such phenomenon will result in some superfluous frequency composition on the frequency spectrum, which will puzzle the detec tion of mechanical vibration. if the sampling frequency is selected from a suitable range, the aliasing phenomenon will be avoided ; the last one is that aliasing frequencies will be produced in zoom demodulation analysis because this algorithm cannot employ digital low - pass filtering to avert the folding frequencies of higher harmonics in the process of zoom sub - sampling

    現有的解調分析方法存在以下三種局限性：將不包括調制信息（故障信息）的兩時域相加信號，也以其頻率之差作為解調信號而解出；廣義檢波濾波解調分析中，由於取絕對值、檢波或平方過程可能產生混頻效應，在解調譜中表現為無法分析的頻率成分，並由此推導出避免這種混頻現象的采樣頻率的選取范圍，從根本上避免此類誤診斷的產生；幾種細化解調分析新演算法中，因為無法在細化分析的選抽時進行數字低通濾波，有可能會出現調制頻率的高次諧波成分發生頻率混疊而反折到低頻部分的現象。
  9. The article gives the results of rectangular moderate - thick plates under various length - width ratios and various thickness - width ratios and the results are compared with corresponding series solution about thin plates. at last the article gives a rough limitation between thin and thick plates according to the absolute values of their relative er -

    本文最後計算了彈性地基上不同長寬比的中厚板在不同的厚寬比情況下的中點撓度值，並與薄板理論的級數精確解答進行比較，根據其相對的誤差絕對值，對彈性地基上薄板和中厚板嘗試進行了大致的界定。
  10. With range of r. v. represents the error limit, accuracy or uncertainty of a measurement, the corresponding probability represents the confidence level of the error limit or uncertainty

    若機變數絕對值小於等於一特定范圍代表相對於一量測平均值之誤差范圍，則對應之機率即代表該量測誤差范圍之信心度。
  11. It is firstly suppose that each node may correspondingly match curve ' s original point. secondly, the curve to match is divide into some sections. the absolute errors between two matching curves are compare in every section

    文中首先假設曲線中的每個節點都可能是待匹配曲線的起點，然後根據等長度分段，比較兩匹配曲線之間的誤差絕對值，由此判斷曲線是否相似。
  12. The calibration of flow model is acceptable with average rms of 0. 7m, residual mean of - 0. 045 m, average absolute mean error of 0. 1 m and normalized rms value of 2. 3 %. the contour map of the simulated heads, elaborated acceptable model calibration compared to observed heads map

    模型結果中，均方差為0 . 7m ，平均誤差為- 0 . 045m ，平均絕對誤差為0 . 1m ，標準均方差為2 . 3 ，模擬地下水流場與實際觀測地下水流場基本一致，說明所建立的數值模型符合該地區的實際水文地質條件。
  13. One is combined with the maximum average error criterion and the mean square error criterion, and the other is combined with the absolute value average error criterion and the mean square error criterion. theory analysis shows that the two new search methods have less computational complexity than that of the non - optimal method

    提出了兩種替換原有均方誤差最小準測的方法，分別是最大平均誤差與均方誤差最小準則相結合的搜索標準，以及絕對值平均誤差與均方誤差最小準則相結合的搜索標準。
分享友人
分享到 Line

分享到臉書