statistics-lab™ 为您的留学生涯保驾护航 在代写理论力学theoretical mechanics方面已经树立了自己的口碑, 保证靠谱, 高质且原创的统计Statistics代写服务。我们的专家在代写理论力学theoretical mechanics代写方面经验极为丰富，各种代写理论力学theoretical mechanics相关的作业也就用不着说。

• Statistical Inference 统计推断
• Statistical Computing 统计计算
• (Generalized) Linear Models 广义线性模型
• Statistical Machine Learning 统计机器学习
• Longitudinal Data Analysis 纵向数据分析
• Foundations of Data Science 数据科学基础
assignmentutor™您的专属作业导师

## 物理代写|理论力学代写theoretical mechanics代考|Error Propagation

We now start from $K$ physical quantities $x_{k}$ with average errors $\sigma_{k}$ and consider the derived quantity $y=f\left(x_{1}, \ldots, x_{K}\right)$. Here all the quantities $x_{k}$ will be independent of each other. What is then the average error in $y$ ?
To begin with, the error $\varepsilon$ in $f\left(x_{1}, \ldots, x_{K}\right)$ is to first order
$$\varepsilon=\sum_{k=1}^{K} \frac{\partial f}{\partial x_{k}} \varepsilon_{k}$$
and hence
\begin{aligned} \sigma^{2} &=\left\langle\varepsilon^{2}\right\rangle \ &=\int \cdots \int_{-\infty}^{\infty}\left(\sum_{k=1}^{K} \frac{\partial f}{\partial x_{k}} \varepsilon_{k}\right)^{2} \rho\left(\varepsilon_{1}, \ldots, \varepsilon_{K}\right) \mathrm{d} \varepsilon_{1} \cdots \mathrm{d} \varepsilon_{K} \ &=\left\langle\sum_{k=1}^{K} \frac{\partial f}{\partial x_{k}} \varepsilon_{k} \cdot \sum_{l=1}^{K} \frac{\partial f}{\partial x_{l}} \varepsilon_{l}\right\rangle=\sum_{k, l=1}^{K} \frac{\partial f}{\partial x_{k}} \frac{\partial f}{\partial x_{l}}\left\langle\varepsilon_{k} \varepsilon_{l}\right\rangle . \end{aligned}
Since the quantities $x_{k}$ and $x_{l}$ should not depend upon each other, they are not correlated to each other (the property $x_{l}$ does not care how large $x_{k}$ is – correlations will be investigated in more detail in Sect. 6.1.5). With $\left\langle\varepsilon_{k}^{2}\right\rangle=\sigma_{k}^{2}$ this leads to
$$\left\langle\varepsilon_{k} \varepsilon_{l}\right\rangle=\left{\begin{array}{cc} \left\langle\varepsilon_{k}\right\rangle\left\langle\varepsilon_{l}\right\rangle & \text { for } k \neq l, \ \sigma_{k}{ }^{2} & \text { for } k=l . \end{array}\right.$$
Here, $\left\langle\varepsilon_{k}\right\rangle=0$ holds for all $k$ (and $l$ ). Therefore, the law of error propagation follows:
$$\sigma^{2}=\sum_{k=1}^{K}\left(\frac{\partial f}{\partial x_{k}}\right)^{2} \sigma_{k}^{2} .$$
In the proof, no normally distributed errors were necessary -thus other distributions with the properties $\left\langle\varepsilon_{k}^{0}\right\rangle=1,\left\langle\varepsilon_{k}^{1}\right\rangle=0$, and $\left\langle\varepsilon_{k}^{2}\right\rangle=\sigma_{k}^{2}$ deliver the error propagation law and with it the basis for all further proofs in this section. In particular, we may invoke this law for repeated measurements of the same quantity, as we shall now do.

## 物理代写|理论力学代写theoretical mechanics代考|Finite Measurement Series and Their Average Error

If we consider the expression
$$\langle x\rangle \approx \bar{x} \equiv \frac{1}{N} \sum_{n=1}^{N} x_{n}$$
as $\bar{x}=f\left(x_{1}, \ldots, x_{N}\right)$, then we can use it in the law of error propagation and deduce that $\partial f / \partial x_{n}=N^{-1}$. Hence, all single measurements enter into the error estimate with the same weight-as already for the estimated value $x_{0}$.

In order to determine the error $\sigma_{n}$, we think of an average over several measurement series, each with $N$ measurements. In this way, we can introduce the average error of the single measurement and find that all single measurements have the same average error $\Delta x$. Therefore, the law of error propagation for $N$ equal terms $N^{-2}(\Delta x)^{2}$ delivers
$$(\Delta \bar{x})^{2}=N \cdot N^{-2}(\Delta x)^{2}=\frac{(\Delta x)^{2}}{N}$$
The average error $\Delta \bar{x}$ in the mean value of the measurement series is thus the $\sqrt{N}$ th part of the average error in a single measurement: the more often measurements are made, the more accurate is the determination of the mean value. However, because of the square root factors, the accuracy can be increased only rather slowly.

Since we do not know the true value $x_{0}$ itself, but only its approximation $\bar{x}$, we still have to account for its uncertainty $\Delta \bar{x}$ in order to determine the average error of the single measurement:
$$(\Delta x)^{2}=\overline{\left(x-x_{0}\right)^{2}}=\overline{\left(x-\bar{x}+\bar{x}-x_{0}\right)^{2}}=\overline{(x-\bar{x})^{2}}+2 \overline{(x-\bar{x})}\left(\bar{x}-x_{0}\right)+\left(\bar{x}-x_{0}\right)^{2} .$$
Here, $\overline{x-\bar{x}}=\bar{x}-\bar{x}=0$ and thus $\overline{(x-\bar{x})^{2}}=\overline{x^{2}}-\bar{x}^{2}$ is rather easy to evaluate. For $\left(\bar{x}-x_{0}\right)^{2}$, we take $(\Delta \bar{x})^{2}=(\Delta x)^{2} / N$. Hence, because $1-N^{-1}=N^{-1}(N-1)$, the average error of the single measurement is
$$(\Delta x)^{2}=\frac{N}{N-1} \overline{(x-\bar{x})^{2}},$$
as claimed previously (see p. 46). And so we have the announced proof. For sufficiently large $N$, we may write $(\Delta x)^{2}=\overline{x^{2}}-\bar{x}^{2}$. The expression $\Delta x$ is referred to as the uncertainty of $x$ in quantum theory (see p. 275).

## 物理代写|理论力学代写theoretical mechanics代考|Error Analysis

How should we modify the result ohtained so far if the same quantity is measured in different ways: first as $x_{1} \pm \Delta x_{1}$, then as $x_{2} \pm \Delta x_{2}$, and so on? What is then the most probable value for $x_{0}$, and what average error does it have?

If the readings of the measurement were taken with the same instrument and equally carefully, the difference in the average errors stems from values $x_{n}$ from measurement series of different lengths. According to the last section, the average error of every single measurement in such a measurement series should be equal to $\Delta x_{n} \sqrt{N_{n}}$, and this independently of $n$ in each of the measurement series. Therefore, the mentioned values $x_{n}$ should contribute with the weight
$$\rho_{n}=\frac{N_{n}}{\sum_{k} N_{k}}=\frac{1}{\left(\Delta x_{n}\right)^{2}} / \sum_{k} \frac{1}{\left(\Delta x_{k}\right)^{2}},$$
whence $\bar{x}=\sum_{n} \rho_{n} x_{n}$ is the properly weighted mean value. The error propagation law delivers
\begin{aligned} \sigma^{2} &=\sum_{n} \rho_{n}^{2} \sigma_{n}^{2}=\frac{1}{\left(\sum_{k}\left(\Delta x_{k}\right)^{-2}\right)^{2}} \sum_{n} \frac{1}{\left(\Delta x_{n}\right)^{4}}\left(\Delta x_{n}\right)^{2}=\frac{\sum_{n}\left(\Delta x_{n}\right)^{-2}}{\left(\sum_{k}\left(\Delta x_{k}\right)^{-2}\right)^{2}} \ &=\frac{1}{\sum_{n}\left(\Delta x_{n}\right)^{-2}} \quad \Longrightarrow \quad \frac{1}{(\Delta x)^{2}}=\sum_{n} \frac{1}{\left(\Delta x_{n}\right)^{2}} \end{aligned}
The more detailed the readings of the measurement, the more important they are for the mean value and for the (un)certainty of the results. These considerations are only then valid without restriction, if the values are compatible with each other within their error limits. If they lie further apart from each other, then we have to take
$$(\Delta x)^{2}=\frac{1}{N-1} \frac{1}{\sum_{n}\left(\Delta x_{n}\right)^{-2}} \sum_{n} \frac{\left(x_{n}-\bar{x}\right)^{2}}{\left(\Delta x_{n}\right)^{2}}$$
Note that, if the values $x_{n}$ do not lie within the error limits, then systematic errors may be involved.

Thus, these two equations answer the questions raised in the general case, where measurements are taken with different instruments and different levels of care: to each value $x_{n}$, we must attach the relative weight $1 /\left(\Delta x_{n}\right)^{2}$.

## 物理代写|理论力学代写theoretical mechanics代考|Error Propagation

$$\varepsilon=\sum_{k=1}^{K} \frac{\partial f}{\partial x_{k}} \varepsilon_{k}$$

$$\sigma^{2}=\left\langle\varepsilon^{2}\right\rangle \quad=\cdots \int_{-\infty}^{\infty}\left(\sum_{k=1}^{K} \frac{\partial f}{\partial x_{k}} \varepsilon_{k}\right)^{2} \rho\left(\varepsilon_{1}, \ldots, \varepsilon_{K}\right) \mathrm{d} \varepsilon_{1} \cdots \mathrm{d} \varepsilon_{K}=\left\langle\sum_{k=1}^{K} \frac{\partial f}{\partial x_{k}} \varepsilon_{k} \cdot \sum_{l=1}^{K} \frac{\partial f}{\partial x_{l}} \varepsilon_{l}\right\rangle=\sum_{k, l=1}^{K} \frac{\partial f}{\partial x_{k}} \frac{\partial f}{\partial x_{l}}\left\langle\varepsilon_{k} \varepsilon_{l}\right\rangle$$

$\$ \$$\mathrm{~ V e f t ~ \ l a n g l e \ v a r e p s i l o n _ { k } ~}$$
\left\langle\varepsilon_{k}\right\rangle\left\langle\varepsilon_{l}\right\rangle \quad \text { for } k \neq l, \sigma_{k}^{2} \quad \text { for } k=l .
$$正确的。$$
\text { Here, } \$\left\langle\varepsilon_{k}\right\rangle=0 \$ \text { holds forall } \$k \$(\text { and } \$1 \$) \text {. There fore, thelawo ferrorpropagation follows : }
$$\backslash sigma^ {2}=\backslash sum_{ {k=1}^{\wedge}{K} \backslash eft \left(\backslash\right. frac {\backslash partial f } \wedge partial \left.x_{-}{k}\right} \backslash right )^{\wedge}{2} \backslash sigma_{k}^{{2} 。 \ \$$

## 物理代写|理论力学代写theoretical mechanics代考|Finite Measurement Series and Their Average Error

$$\langle x\rangle \approx \bar{x} \equiv \frac{1}{N} \sum_{n=1}^{N} x_{n}$$

$$(\Delta \bar{x})^{2}=N \cdot N^{-2}(\Delta x)^{2}=\frac{(\Delta x)^{2}}{N}$$

$$(\Delta x)^{2}=\overline{\left(x-x_{0}\right)^{2}}=\overline{\left(x-\bar{x}+\bar{x}-x_{0}\right)^{2}}=\overline{(x-\bar{x})^{2}}+2 \overline{(x-\bar{x})}\left(\bar{x}-x_{0}\right)+\left(\bar{x}-x_{0}\right)^{2} .$$

$$(\Delta x)^{2}=\frac{N}{N-1} \overline{(x-\bar{x})^{2}}$$

## 物理代写|理论力学代写theoretical mechanics代考|Error Analysis

$$\rho_{n}=\frac{N_{n}}{\sum_{k} N_{k}}=\frac{1}{\left(\Delta x_{n}\right)^{2}} / \sum_{k} \frac{1}{\left(\Delta x_{k}\right)^{2}},$$

$$\sigma^{2}=\sum_{n} \rho_{n}^{2} \sigma_{n}^{2}=\frac{1}{\left(\sum_{k}\left(\Delta x_{k}\right)^{-2}\right)^{2}} \sum_{n} \frac{1}{\left(\Delta x_{n}\right)^{4}}\left(\Delta x_{n}\right)^{2}=\frac{\sum_{n}\left(\Delta x_{n}\right)^{-2}}{\left(\sum_{k}\left(\Delta x_{k}\right)^{-2}\right)^{2}} \quad=\frac{1}{\sum_{n}\left(\Delta x_{n}\right)^{-2}} \quad \Longrightarrow \quad \frac{1}{(\Delta x)^{2}}=\sum_{n} \frac{1}{\left(\Delta x_{n}\right)^{2}}$$

$$(\Delta x)^{2}=\frac{1}{N-1} \frac{1}{\sum_{n}\left(\Delta x_{n}\right)^{-2}} \sum_{n} \frac{\left(x_{n}-\bar{x}\right)^{2}}{\left(\Delta x_{n}\right)^{2}}$$

## 有限元方法代写

assignmentutor™作为专业的留学生服务机构，多年来已为美国、英国、加拿大、澳洲等留学热门地的学生提供专业的学术服务，包括但不限于Essay代写，Assignment代写，Dissertation代写，Report代写，小组作业代写，Proposal代写，Paper代写，Presentation代写，计算机作业代写，论文修改和润色，网课代做，exam代考等等。写作范围涵盖高中，本科，研究生等海外留学全阶段，辐射金融，经济学，会计学，审计学，管理学等全球99%专业科目。写作团队既有专业英语母语作者，也有海外名校硕博留学生，每位写作老师都拥有过硬的语言能力，专业的学科背景和学术写作经验。我们承诺100%原创，100%专业，100%准时，100%满意。

## MATLAB代写

MATLAB 是一种用于技术计算的高性能语言。它将计算、可视化和编程集成在一个易于使用的环境中，其中问题和解决方案以熟悉的数学符号表示。典型用途包括：数学和计算算法开发建模、仿真和原型制作数据分析、探索和可视化科学和工程图形应用程序开发，包括图形用户界面构建MATLAB 是一个交互式系统，其基本数据元素是一个不需要维度的数组。这使您可以解决许多技术计算问题，尤其是那些具有矩阵和向量公式的问题，而只需用 C 或 Fortran 等标量非交互式语言编写程序所需的时间的一小部分。MATLAB 名称代表矩阵实验室。MATLAB 最初的编写目的是提供对由 LINPACK 和 EISPACK 项目开发的矩阵软件的轻松访问，这两个项目共同代表了矩阵计算软件的最新技术。MATLAB 经过多年的发展，得到了许多用户的投入。在大学环境中，它是数学、工程和科学入门和高级课程的标准教学工具。在工业领域，MATLAB 是高效研究、开发和分析的首选工具。MATLAB 具有一系列称为工具箱的特定于应用程序的解决方案。对于大多数 MATLAB 用户来说非常重要，工具箱允许您学习应用专业技术。工具箱是 MATLAB 函数（M 文件）的综合集合，可扩展 MATLAB 环境以解决特定类别的问题。可用工具箱的领域包括信号处理、控制系统、神经网络、模糊逻辑、小波、仿真等。