### 统计代写|机器学习代写machine learning代考|The ν–Trick and Other Variants

statistics-lab™ 为您的留学生涯保驾护航 在代写机器学习machine learning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。我们的专家在代写机器学习machine learning代写方面经验极为丰富，各种代写机器学习machine learning相关的作业也就用不着说。

• Statistical Inference 统计推断
• Statistical Computing 统计计算
• Advanced Probability Theory 高等概率论
• Advanced Mathematical Statistics 高等数理统计学
• (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

## 统计代写|机器学习代写machine learning代考|Multiclass Support Vector Machines

In order to extend the SV learning algorithm to $K=|\mathcal{Y}|>2$ classes two different strategies have been suggested.

1. The first method is to learn $K$ SV classifiers $f_{j}$ by labeling all training points having $y_{i}=j$ with $+1$ and $y_{i} \neq j$ with $-1$ during the training of the $j$ th classifier. In the test stage, the final decision is obtained by
$f_{\text {multiple }}(x)=\underset{y \in \mathcal{Y}}{\operatorname{argmax}} f_{y}(x) .$
Clearly, this method learns one classifier for each of the $K$ classes against all the other classes and is hence known as the one-versus-rest (o-v-r) method. It can be

shown that it is possible to solve the $K$ optimization problems at once. Note that the computational effort is of order $\mathcal{O}\left(K m^{2}\right)$.

1. The second method is to learn $K(K-1) / 2 \mathrm{SV}$ classifiers. If $1 \leq i<j \leq K$ the classifiers $f_{i, j}$ is learned using only the training samples from the class $i$ and $j$, labeling them $+1$ and $-1$, respectively. This method has become known as the one-versus-one (o-v-o) method. Given a new test object $x \in \mathcal{X}$, the frequency $n_{i}$ of “wins” for class $i$ is computed by applying $f_{i, j}$ for all $j$. This results in a vector $\mathbf{n}=\left(n_{1} ; \ldots ; n_{K}\right)$ of frequencies of “wins” of each class. The final decision is made for the most frequent class, i.e.,
$f_{\text {multiple }}(x)=\underset{y \in \mathcal{Y}}{\operatorname{argmax}} n_{y}$
Using a probabilistic model for the frequencies $\mathbf{n}$, different prior probabilities of the classes $y \in \mathcal{Y}$ can be incorporated, resulting in better generalization ability. Instead of solving $K(K-1) / 2$ separate optimization problems, it is again possible to combine them in a single optimization problem. If the prior probabilities $\mathbf{P}_{\mathbf{Y}}(j)$ for the $K$ classes are roughly $\frac{1}{K}$, the method scales as $\mathcal{O}\left(m^{2}\right)$ and is independent of the number of classes.

Recently, a different method for combining the single pairwise decisions has been suggested. By specifying a directed acyclic graph (DAG) of consecutive pairwise classifications, it is possible to introduce a class hierarchy. The leaves of such a DAG contain the final decisions which are obtained by exclusion rather than by voting. This method compares favorably with the $o-v-o$ and $o-v-r$ methods.

## 统计代写|机器学习代写machine learning代考|Support Vector Regression Estimation

In the regression estimation problem we are given a sample of $m$ real target values $\boldsymbol{t}=\left(t_{1}, \ldots, t_{m}\right) \in \mathbb{R}^{m}$, rather than $m$ classes $\boldsymbol{y}=\left(y_{1}, \ldots, y_{m}\right) \in \mathcal{Y}^{m} .$ In order to extend the SV learning algorithm to this task, we note that an “inversion” of the linear loss $l_{\text {lin }}$ suffices in order to use the $S V$ machinery for real-valued outputs $t_{i}$. In classification the linear loss $l_{\operatorname{lin}}(f(x), \cdot)$ adds to the total cost, if the real-valued output of $|f(x)|$ is smaller than 1. For regression estimation it is desirable to have the opposite true, i.e., incurred costs result if $|t-f(x)|$ is very large instead of small. This requirement is formally captured by the $\varepsilon-i n$ sensitive loss
$$l_{\varepsilon}(f(x), t)= \begin{cases}0 & \text { if }|t-f(x)| \leq \varepsilon \ |t-f(x)|-\varepsilon & \text { if }|t-f(x)|>\varepsilon\end{cases}$$ Then, one obtains a quadratic programming problem similar to $(2.46)$, this time in $2 m$ dual variables $\alpha_{i}$ and $\tilde{\alpha}{i}$-two corresponding to each training point constraint. This is simply due to the fact that $f$ can fail to attain a deviation less than $\varepsilon$ on both sides of the given real-valued output $t{i}$, i.e., $t_{i}-\varepsilon$ and $t_{i}+\varepsilon$. An appealing feature of this loss is that it leads to sparse solutions, i.e., only a few of the $\alpha_{i}$ (or $\tilde{\alpha}_{i}$ ) are non-zero. For further references that cover the regression estimation problem the interested reader is referred to Section 2.6.

## 统计代写|机器学习代写machine learning代考|Support Vector Machines for Classification

A major drawback of the soft margin SV learning algorithm given in the form (2.48) is the lack of control over how many training points will be considered as margin errors or “outliers”, that is, how many have $\tilde{\gamma}{i}\left(w{\mathrm{SVM}}\right)<1$. This is essentially due to the fact that we fixed the functional margin to one. By a simple reparameterization it is possible to make the functional margin itself a variable of the optimization problem. One can show that the solution of the following optimization problem has the property that the new parameter $v$ bounds the fraction of margin errors $\frac{1}{m}\left|\left{\left(x_{i}, y_{i}\right) \in z \mid \tilde{\gamma}{i}\left(\mathbf{w}{\text {SVM }}\right)<\rho\right}\right|$ from above:
$$\begin{array}{ll} \operatorname{minimize} \quad & \frac{1}{m} \sum_{i=1}^{m} \xi_{i}-v \rho+\frac{1}{2}|\mathbf{w}|^{2} \ \text { subject to } & y_{i}\left\langle\mathbf{x}{i}, \mathbf{w}\right\rangle \geq \rho-\xi{i} \quad i=1, \ldots, m, \ & \boldsymbol{\xi} \geq \mathbf{0}, \rho \geq 0 . \end{array}$$
It can be shown that, for each value of $v \in[0,1]$, there exists a value of $\lambda \in \mathbb{R}^{+}$ such that the solution $\mathbf{w}{\nu}$ and $\mathbf{w}{\lambda}$ found by solving $(2.52)$ and (2.48) have the same geometrical margins $\gamma_{z}\left(\mathbf{w}{\nu}\right)=\gamma{z}\left(\mathbf{w}{\lambda}\right)$. Thus we could try different values of $\lambda$ in the standard linear soft margin SVM to obtain a required fraction of margin errors. The appealing property of the problem $(2.52)$ is that this adjustment is done within the one optimization problem (see Section B.5). Another property which can be proved is that, for all probability models where neither $\mathbf{P}{X}({X, 1})$ nor $\mathbf{P}{\mathbf{X}}({\mathrm{X},-1})$ contains any discrete component, $v$ asymptotically equals the fraction of margin errors. Hence, we can incorporate prior knowledge of the noise level $\mathbf{E}{\mathrm{X}}\left[1-\max {y \in \mathcal{Y}}\left(\mathbf{P}{\mathrm{Y} \mid \mathrm{X}=x}(y)\right)\right]$ via $v$. Excluding all training points for which the real-valued output is less than $\rho$ in absolute value, the geometrical margin of the solution on the remaining training points is $\rho /|\mathbf{w}|$.

## 统计代写|机器学习代写machine learning代考|Multiclass Support Vector Machines

1. 第一种方法是学习ķSV分类器Fj通过标记所有训练点是一世=j和+1和是一世≠j和−1在训练期间j分类器。在测试阶段，最终的决定是通过
F多 (X)=最大参数是∈是F是(X).
显然，该方法为每个ķ类对所有其他类，因此被称为一对休息（ovr）方法。有可能

1. 第二种方法是学习ķ(ķ−1)/2小号在分类器。如果1≤一世<j≤ķ分类器F一世,j仅使用类中的训练样本学习一世和j, 标记它们+1和−1， 分别。这种方法已被称为一对一（ovo）方法。给定一个新的测试对象X∈X, 频率n一世课堂上的“胜利”一世通过应用计算F一世,j对全部j. 这导致一个向量n=(n1;…;nķ)每个类别的“获胜”频率。最终决定是为最频繁的类做出的，即
F多 (X)=最大参数是∈是n是
使用频率的概率模型n, 类的不同先验概率是∈是可以合并，从而获得更好的泛化能力。而不是解决ķ(ķ−1)/2单独的优化问题，可以再次将它们组合成一个优化问题。如果先验概率磷是(j)为了ķ类大致1ķ, 该方法缩放为这(米2)并且与类的数量无关。

## 统计代写|机器学习代写machine learning代考|Support Vector Regression Estimation

le(F(X),吨)={0 如果 |吨−F(X)|≤e |吨−F(X)|−e 如果 |吨−F(X)|>e然后，得到一个二次规划问题，类似于(2.46), 这次在2米对偶变量一种一世和一种~一世-两个对应于每个训练点约束。这仅仅是因为F可能无法获得小于以下的偏差e在给定实值输出的两边吨一世， IE，吨一世−e和吨一世+e. 这种损失的一个吸引人的特点是它会导致稀疏的解决方案，即只有少数一种一世（或者一种~一世) 非零。有关回归估计问题的更多参考资料，感兴趣的读者请参阅第 2.6 节。

## 有限元方法代写

tatistics-lab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。