基于菌群优化算法和小波SVM的P2P流量识别方法

发布时间：2018-04-14 19:40

本文选题：P2P + 流量识别　；参考：《湖北工业大学》2014年硕士论文

【摘要】：由于互联网本身是基于点到点的传输，使得P2P技术得到了广泛地应用与发展，给用户带来便捷的同时也给网络质量和网络管理带来了巨大的负面影响，如网络拥堵、知识产权、资源管理以及安全隐患问题。P2P流量识别问题得到了越来越多的关注，，围绕P2P识别问题产生了一批相关的算法。近年来研究与应用最为广泛的P2P流量识别方法之一是基于机器学习的识别方法。然而由于P2P网络的突变性和不确定性，对于传统的基于贝叶斯网络，决策树算法等机器学习方法而言，P2P流量识别变得更加困难。支持向量机（Support Vector Machine，简称SVM）作为目前性能良好且广泛使用的分类器，它在克服“维数灾难”以及避免局部最优解等流量识别问题上具有明显的优势。然而在P2P流量识别问题中，支持向量机的性能受惩罚系数和核函数的参数的影响。常规的支持向量机参数求取方法性能需要进一步加强。菌群优化算法是近年来新提出的一种群体智能优化算法，具有较强的寻优能力。小波函数可以用来描述突变信号逐渐精细的特点，能够在一定程度上处理突变的P2P网络。因此本文重点研究菌群优化算法和小波支持向量机在P2P流量识别问题中的应用，本文的主要研究内容和工作如下： 1提出了一种结合菌群优化算法和支持向量机的P2P流量识别方法。首先引入菌群优化算法来优化支持向量机的两个参数，从而可以得到参数配置较优的支持向量机；并将其应用于P2P流量识别。通过与现有的基于遗传算法优化参数和基于粒子群算法优化参数的支持向量机方法在实际的P2P流量识别问题中进行性能对比测试，结果显示基于菌群优化算法所优化的支持向量机在性能上更具优势。 2在优化了支持向量机的参数以后，对配置不同核函数的支持向量机进行P2P流量识别性能测试和分析。由于小波分析在局部分析和处理突变信号方面的优越性，能很好地解决P2P网络流量的突变性和不确定性，这里重点研究选择合适的小波核函数来提高支持向量机的性能，通过对常用的核函数及多种小波核函数的对比实验，结果表明基于小波核函数和基于菌群优化算法的SVM在P2P流量识别具有较高的识别精度与稳定性。
[Abstract]:Since the Internet itself is based on point-to-point transmission, P2P technology has been widely used and developed, which brings convenience to users, but also brings huge negative effects to network quality and network management, such as network congestion, intellectual property rights.More and more attention has been paid to resource management and P2P traffic identification.In recent years, one of the most widely studied and applied P2P traffic identification methods is based on machine learning.However, due to the mutation and uncertainty of P2P network, it is more difficult for traditional machine learning methods based on Bayesian network and decision tree algorithm to identify P2P traffic.As a widely used classifier with good performance at present, support Vector Machine (SVM) has obvious advantages in overcoming "dimension disaster" and avoiding the problem of local optimal solution.However, in P2P traffic identification problem, the performance of SVM is affected by the penalty coefficient and the parameters of kernel function.The performance of the conventional support vector machine (SVM) parameter estimation method needs to be further enhanced.Bacterial colony optimization algorithm is a new swarm intelligence optimization algorithm proposed in recent years.Wavelet function can be used to describe the characteristics of the gradual refinement of abrupt signals and to deal with sudden changes in P2P networks to a certain extent.Therefore, this paper focuses on the application of microbial colony optimization algorithm and wavelet support vector machine in P2P traffic identification. The main contents and work of this paper are as follows:1. A P2P traffic identification method combined with microbial colony optimization algorithm and support vector machine is proposed.Firstly, the bacterial colony optimization algorithm is introduced to optimize the two parameters of support vector machine (SVM), so that the support vector machine with better parameter configuration can be obtained, and it is applied to P2P traffic identification.Compared with the existing support vector machine (SVM) based on genetic algorithm and particle swarm optimization parameters, the performance of P2P traffic identification problem is compared and tested.The results show that the support vector machine based on colony optimization algorithm has more advantages in performance.2 after optimizing the parameters of support vector machine, the P2P traffic identification performance of support vector machine with different kernel function is tested and analyzed.Due to the advantages of wavelet analysis in local analysis and processing of abrupt signals, the mutation and uncertainty of P2P network traffic can be solved well. This paper focuses on the selection of appropriate wavelet kernel functions to improve the performance of support vector machines.The comparative experiments of common nuclear function and several wavelet kernel functions show that SVM based on wavelet kernel function and bacterial colony optimization algorithm has high recognition accuracy and stability in P2P traffic identification.
【学位授予单位】：湖北工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP181;TP393.06

【参考文献】