基于数据挖掘的移动设备使用情况分析的研究

发布时间：2018-02-07 11:56

本文关键词： 数据挖掘 K-means聚类算法移动设备 Apriori关联规则算法　出处：《内蒙古大学》2017年硕士论文　论文类型：学位论文

【摘要】：近年来,在互联网的飞速发展下,手机移动设备也从3G通信时代进入现在的4G互联时代。在2020年,我国预计将大规模运用5G移动通信技术。然而,当今数据挖掘领域的一个热点话题便是对移动应用设备产生的丰富数据进行分析与挖掘。本篇论文将引入数据挖掘技术来研究移动设备的使用情况并进行分析,最终为移动设备的运营商提供一定的参考。本文的研究工作主要包括三部分:第一部分:对本实验数据文件进行整合,对整合好的数据文件利用Excel进行整体的分析,从用户基本特性、时间特性、手机品牌特性、地理位置特性、使用的手机App种类这几方面进行研究,找到其中有价值的研究点。第二部分:从宏观上,运用数据挖掘中的聚类算法对经纬度数据进行聚类,通过使用Python语言编程来实现算法的功能并对聚类结果进行可视化。观察聚类后的结果,可以看到呈现出的结果大致是一个中国地图的形状,并且在中国地图的轮廓之外还存在一些较少的分布。重新设置k值对数据进行二次聚类,调用百度地图的API接口将二次聚类结果的每一个簇的质心数据进行图形化显示,并观察移动设备在地图上各个地方的分布情况,从而进行具体的分析。第三部分:从微观上,利用数据挖掘技术对移动设备用户使用的手机App种类进行关联规则挖掘,借助数据挖掘软件Weka中自带的Apriori算法进行关联规则挖掘,将挖掘出的实验结果进行分析。
[Abstract]:In recent years, with the rapid development of the Internet, mobile devices have also entered the 4G interconnection era from the 3G communication era. In 2020, China is expected to use 5G mobile communication technology on a large scale. However, Nowadays, a hot topic in the field of data mining is to analyze and mine the rich data generated by mobile application devices. This paper will introduce data mining technology to study the use of mobile devices and analyze them. The research work of this paper mainly includes three parts: the first part: the integration of the experimental data file, the integration of the integrated data file using Excel for the overall analysis, From the user's basic characteristics, time characteristics, mobile phone brand characteristics, geographical location characteristics, the use of mobile phone App types of these aspects to find the valuable research point. Part two: from the macro perspective, The longitude and latitude data are clustered by using the clustering algorithm in data mining, and the function of the algorithm is realized by programming with Python language, and the result of clustering is visualized. You can see that the result is roughly the shape of a Chinese map, and there are fewer distributions outside the outline of the Chinese map. The API interface of Baidu Maps is used to graphically display the centroid data of each cluster, and to observe the distribution of mobile devices on the map. The association rules of mobile phone App used by mobile device users are mined by using data mining technology. The association rules are mined with the help of the Apriori algorithm included in the data mining software Weka, and the experimental results are analyzed.
【学位授予单位】：内蒙古大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【相似文献】