蜂窝无线网络中聚类分析与资源优化系统的设计与实现

发布时间：2018-05-04 02:24

本文选题：蜂窝无线网络 + 聚类分析　；参考：《北京邮电大学》2016年硕士论文

【摘要】：随着LTE技术的高速发展与广泛应用,移动用户的行为规律也随之发生了很大变化。无线技术的飞速发展,带来的是移动终端使用频率增大,使用时长变多。而无线数据的爆发增长,带来的是日益增多的运营机遇和数据业务,与此同时,移动端数据的增长趋势使得运营商原本对网络资源优化难度不断增大,因而采用新的分析方法,即同时分析用户、业务、资源多维度的数据对于无线网络规划和资源优化有重大意义。LTE网络承载移动互联网数据业务的逐步发展,为了应对数据种类和数量激增的海量通信数据,本文将数据挖掘的分析方法应用到蜂窝无线网络数据处理与分析中。另一方面,用户通过移动设备接入互联网的频率日益增长,搭载安卓操作系统的设备的占有率超过80%,安卓系统具有极好的代表性。基于以上背景,本论文首先对蜂窝无线网络中的用户、业务、资源数据研究分析,并给出一种基于聚类分析的多维度拟合关联预测方法。在此之后搭建了一个基于安卓的移动互联网终端分析系统,实现了从用户端采集数据,并使用聚类算法分析,将结果展示在web端的整体流程。本论文主要内容如下:1.给出了一个基于聚类分析的多维度拟合关联预测方法。在该方法中,首先通过主成分分析,将待分析样本数据进行降维处理,随后通过二次聚类的方法对数据聚类分析,并将结果采用梯度拟合,进行蜂窝无线网络数据的关联预测。研究结果表明:与对数据直接拟合相比,聚类分析能够提高对大量LTE数据预测的精度,实现资源预测方面的优化。2.设计了基于安卓的移动互联网终端分析系统。该系统包括基于安卓终端的数据采集系统和服务器端的数据分析系统两部分。在安卓终端的采集系统可以采集蜂窝用户的多种数据,包括用户基本信息、信号数据、通话数据、短信数据、流量业务数据以及通信异常数据等,同时终端记录发生业务时的GPS坐标。服务器端的分析系统存储终端采集的数据,实现本文的聚类算法,并使用现网真实数据进行分析预测。本文设计的分析系统,结合本文给出的方法,为蜂窝无线网络的数据采集、分析,预测提供了一个完整的解决流程。
[Abstract]:With the rapid development and wide application of LTE technology, the behavior of mobile users has changed greatly. With the rapid development of wireless technology, the frequency of mobile terminal is increasing and the time is long. The explosion and growth of wireless data brings more and more operation opportunities and data services. At the same time, the increasing trend of mobile data makes it more difficult for operators to optimize network resources, so a new analysis method is adopted. That is to say, it is of great significance for wireless network planning and resource optimization to analyze the multi-dimensional data of users, services and resources at the same time. LTE network carries the gradual development of mobile Internet data services. In order to deal with the huge amount of communication data, the analysis method of data mining is applied to the data processing and analysis of cellular wireless network. On the other hand, the frequency of users accessing the Internet through mobile devices is increasing, with more than 80% of devices running Android operating system, which is an excellent representative. Based on the above background, this paper first studies and analyzes the data of users, services and resources in cellular wireless networks, and presents a multi-dimensional fitting association prediction method based on clustering analysis. After that, a mobile Internet terminal analysis system based on Android is built, which can collect data from the user side and use clustering algorithm to analyze the result. The result is displayed in the whole flow of web. The main contents of this thesis are as follows: 1. A multi-dimensional fitting association prediction method based on cluster analysis is presented. In this method, firstly, the sample data are reduced by principal component analysis, then the data are clustered by quadratic clustering method, and the results are fitted by gradient to predict the data association of cellular wireless network. The results show that the clustering analysis can improve the precision of forecasting a large number of LTE data and realize the optimization of resource prediction by comparing with the direct fitting of the data. A mobile internet terminal analysis system based on Android is designed. The system includes two parts: the data acquisition system based on Android terminal and the data analysis system on the server side. The collection system of Android terminal can collect a variety of data from cellular users, including user basic information, signal data, call data, short message data, traffic data and abnormal communication data, etc. At the same time, the terminal records the GPS coordinates when the business occurs. The server side analysis system stores the data collected by the terminal, realizes the clustering algorithm of this paper, and uses the real data of the existing network to analyze and predict. The analysis system designed in this paper, combined with the method given in this paper, provides a complete solution for the data acquisition, analysis and prediction of cellular wireless network.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN929.5;TP311.52

【参考文献】