### 数学代写|理论计算机代写theoretical computer science代考|Methodology

statistics-lab™ 为您的留学生涯保驾护航 在代写理论计算机theoretical computer science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计算机theoretical computer science代写方面经验极为丰富，各种代写理论计算机theoretical computer science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

## 数学代写|理论计算机代写theoretical computer science代考|Methodology

Figure 2 illustrates the overall scheme of our methodology. There are three main modules Trace Generator, Language Learner, and Classifier. The trace generator module converts the packet traces of each application into a word list by first splitting the traces by using the timing parameters and then applying the Map function to identified network units. Then, the k-TSS network language of the application is learned from its words by the Language Learner module. By applying this process for all applications, a database of k-TSS languages is obtained. To classify traffic of a system, first, the Trace Generator module extracts its sessions and their network units using the timing parameters. Then, it converts the sessions to the symbolic words to prepare the inputs of the Language Learner module. Finally, the Classifier module compares the learned language of each session of the given traffic against the languages of the database to label them. We describe each module in detail in the following.

## 数学代写|理论计算机代写theoretical computer science代考|Trace Generator

The goal of the Trace Generator module is to transform the packet traces of each application into traces of smaller units, while preserving the relation among these units, as the letters of a language. To do so, at first, packet traces are split into smaller units, named network unit, by using the three timing parameters, introduced in Sect. 3. Then, these units are categorized based on their higherlayer protocol, and each category is clustered separately and named upon it. Consequently, the names of clusters constitute the Network Alphabet. Finally, the network alphabets are put in together to form our traces. Therefore, the input of this subproblem is $\Pi$ containing $m$ application traces and the timing parameters st, $f d$, and it, while the output is a set of word list $S=\left{W_{1}, W_{2}, \ldots, W_{m}\right}$, where $W_{i}$ is a list including the generated words of language $A p p_{i}$.

Network Unit Extraction. Network Units are extracted during three steps. At first, network traces are split into sessions based on the session threshold. Then, flow-sessions are extracted based on the inactive timeort. After that, flowsessions are divided into smaller units according to the flow duration constraint. Figure 1 shows the result of applying this module on a given trace to produce sessions, flow-sessions, and network units.

Network Unit Clustering. For naming network units, there is a need to group the similar ones and name them upon it. As there is no information on the number of units or their labels, an unsupervised machine learning algorithm named Kmeans $+t$ is used for clustering these units. The only information available for each flow are those stored in the frame header of its packets. Therefore, the clustering algorithm could benefit from the knowledge of the unit higher-layer protocol. Consequently, flows having the same protocol are at worst clustered together and named upon their protocol. The packets of each flow can be further clustered in terms of multiple statistical features listed in Table 1. Two groups of features are used for clustering network unit, the first group consists of features number one to six, and the second group includes features number three to nine. Regarding these two groups, the statistical features group (stats) is another configuration of this problem.

Moreover, it is required to set the number of clusters before performing the Kmeanst+ algorithm. To do so, the number of clusters is automatically determined with the help of the elbow [17] method in our approach.

Our defined Map function returns a concatenation of the symbol abbreviating packet protocol name and a natural number indicating its cluster number (to distinguish network units). For example, TL-5 corresponds to a packet sequence belonging to cluster 5 of the TLS protocol.

## 数学代写|理论计算机代写theoretical computer science代考|Language Learner

After preparation tasks, we obtain the modified word list $W^{\prime}=$ $\left[w_{1}^{\prime}, w_{2}^{\prime}, \ldots, w_{n}^{\prime}\right]$. Now, we learn a list of k-TSS languages for each $w^{\prime} \in$ $W^{\prime}$ as $\mathcal{L}\left(w^{\prime}\right)$ by a k-size frame scanner. For instance, for the given $w^{\prime}=$ SL-4 SL-2 T-10 TL-2 T-1 U-2, $\mathcal{L}\left(w^{\prime}\right)$ is obtained by k-TSS vector $Z\left(w^{\prime}\right)$ as $\langle\Sigma={\mathrm{SL}-2, \mathrm{TL}-2, \mathrm{~T}-1, \mathrm{U}-2, \mathrm{SL}-4, \mathrm{~T}-10}, I={\mathrm{SL}-4 \mathrm{SL}-2,}, F=$, ${\mathrm{T}-1 \mathrm{U}-2}, T={\mathrm{SL}-4 \mathrm{SL}-2 \mathrm{~T}-10, \mathrm{SL}-2 \mathrm{~T}-10 \mathrm{TL}-2, \mathrm{~T}-10 \mathrm{TL}-2 \mathrm{~T}-1, \mathrm{TL}-2$ T-1 U-2} .

To specify each word’s class, we observe its segments instead of exploring the whole word to be more noise tolerable. To this aim, we reduce the classification problem to find the similarity between the generated k-TTS languages of the trace and the ones in the database. For each k-TSS language in the database, we calculate its proximity with $\mathcal{L}\left(w^{\prime}\right)$ by a distance function defined as:

Definition 6 (Distance Function). Distance function $D$ measures the proximity metric between two k-TSS languages. Let $Z=\langle\Sigma, I, F, T\rangle$ be the $k$-test vector of $\mathcal{L}\left(w^{\prime}\right)$ and $Z_{i}=\left\langle\Sigma_{i}, I_{i}, F_{i}, T_{i}\right\rangle$ be the k-test vector of $\mathcal{L}\left(A p p_{i}\right)$. Then, $D\left(\mathcal{L}\left(w^{\prime}\right), \mathcal{L}\left(A p p_{i}\right)\right)$ is computed by five auciliary variables, measuring the sets diffenence fraction: $\triangle T, \Delta T_{i}, \Delta \Sigma, \triangle I$ and $\triangle F$ (defined in Eq. 1).
$\Delta T=\frac{T-T_{i}}{T}, \Delta T_{i}=\frac{T_{i}-T}{T_{i}}, \Delta \Sigma=\frac{\Sigma-\Sigma_{i}}{\Sigma}, \Delta I=\frac{I-I_{i}}{I}, \Delta F=\frac{F-F_{i}}{F} .$
$D\left(\mathcal{L}\left(w^{\prime}\right), \mathcal{L}\left(A p p_{i}\right)\right)=\overline{\Delta^{\prime} T} \overline{\triangle^{\prime} T_{i}} \overline{\Delta^{\prime} \Sigma} \overline{\triangle^{\prime} I} \overline{\triangle^{\prime} F}$
We assume a priority among these delta metrics as $\triangle T, \Delta T_{i}, \triangle \Sigma, \triangle I$ and $\triangle F$. Since $T$ carries more information of words than $I$ and $F$, we assign the highest priority to $\triangle T$. Also, we give $\triangle T$ higher priority than $\triangle T_{i}$, as $T$ is generated from a test trace in contrast to $T_{i}$ which is generated from a number of traces of an application. We specify the next priority to $\triangle \Sigma$ to take into account the alphabet sets differences. We assign the next priorities to $\triangle I$ and then $\triangle F$, respectively, because the trimming operation of the preparation phase may impact on the end of the words while their initials are not modified. Finally, we convert the value of delta metrics from a float number in the range $[0,1]$ to an integer number in the range $[0,99]$, renaming them by a prime sign, for example, $\triangle^{\prime} T$. By this priority, the distance function is defined as given by Eq. 2 .

A word $w$ is categorized in class $j$ if the k-TSS language of $w$ has the minimum distance with the k-TSS language of $A p p_{j}$ among all the applications:
$$\operatorname{Class}\left(w^{\prime}\right)=j \Longleftrightarrow \text { if } D\left(\mathcal{L}\left(w^{\prime}\right), \mathcal{L}\left(A p p_{j}\right)\right)=\operatorname{argmin}{\forall A p p{i} \in|\mathcal{A}|}\left(D\left(\mathcal{L}\left(w^{\prime}\right), \mathcal{L}\left(A p p_{i}\right)\right)\right)$$
Considering the running example $\left(\mathcal{L}\left(A p p_{i}\right)\right.$ in Sect. $4.2$ and $\left.\mathcal{L}\left(w^{\prime}\right)\right)$, the defined delta metrics are obtained as $\triangle T=0.75\left(\triangle^{\prime} T=74\right), \Delta T_{i}=0.85$ $\left(\triangle^{\prime} T_{i}=84\right), \triangle \Sigma=0.16\left(\triangle^{\prime} \Sigma=16\right), \triangle I=0\left(\triangle^{\prime} I=0\right)$, and $\triangle F=1\left(\triangle^{\prime} F=\right.$ $99)$ and finally, the distance is computed as $D\left(\mathcal{L}\left(w^{\prime}\right), \mathcal{L}\left(A p p_{i}\right)\right)=7484160099 .$

## 数学代写|理论计算机代写theoretical computer science代考|Trace Generator

Trace Generator 模块的目标是将每个应用程序的数据包跟踪转换为较小单元的跟踪，同时将这些单元之间的关系保留为语言的字母。为此，首先，通过使用第 3 节中介绍的三个时间参数，将数据包跟踪分成更小的单元，称为网络单元。3. 然后，这些单元根据它们的高层协议进行分类，每个类别被单独聚类并命名。因此，集群的名称构成了网络字母表。最后，将网络字母组合在一起形成我们的踪迹。因此，这个子问题的输入是圆周率包含米应用程序跟踪和时序参数 st，Fd, 而它, 而输出是一组单词列表S=\left{W_{1}, W_{2}, \ldots, W_{m}\right}S=\left{W_{1}, W_{2}, \ldots, W_{m}\right}， 在哪里在一世是一个列表，包括生成的语言单词一种pp一世.

## 数学代写|理论计算机代写theoretical computer science代考|Language Learner

Δ吨=吨−吨一世吨,Δ吨一世=吨一世−吨吨一世,ΔΣ=Σ−Σ一世Σ,Δ一世=一世−一世一世一世,ΔF=F−F一世F.
D(大号(在′),大号(一种pp一世))=Δ′吨¯△′吨一世¯Δ′Σ¯△′一世¯△′F¯

## 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。