基于智能手机应用数据的用户属性挖掘

发布时间：2018-01-24 20:58

本文关键词： 用户属性数据挖掘应用安装列表应用使用序列　出处：《浙江大学》2017年硕士论文　论文类型：学位论文

【摘要】：人们迫切地随时随地从互联网获取信息和服务,这加快了移动互联网的发展,同时促进了移动互联网快速迈入了"大数据"时代。长期以来,智能手机一直是移动互联网的主体,它为了适应用户爆炸式的需求不断地技术创新。用户频繁地使用智能手机,产生了海量的个性化日志数据。这些数据与用户的行为紧密关联,所以隐含了非常广泛的用户属性,如性别、年龄等个人信息以及用户的兴趣爱好和生活习惯等。通过挖掘用户的各种属性,可以加深对用户全面而客观的理解,进而帮助我们来持续改进设备、服务与应用。近年来,越来越多的人开始使用智能手机相关数据来挖掘用户属性,他们大多基于全球定位系统、基站、无线网热点等位置信息,与呼叫详细记录、通讯录等社交信息进行研究。智能手机上的应用,是用户使用手机各种服务的主要入口,手机应用相关的数据也能反映用户广泛的基本属性、兴趣偏好、生活习惯等,这提供给我们一个新的机会来更好地理解用户。智能手机上的应用安装列表数据和应用使用序列数据非常容易采集,并且蕴含丰富的信息量。本文针对性地从这些手机应用相关数据中挖掘用户的属性,完成了以下工作:1.提出了基于手机应用数据的用户属性挖掘方案,并围绕应用安装列表数据和应用使用序列数据设计出用户属性挖掘的具体框架。2.阐述了全面的数据预处理过程,对两组数据做了全面的剖析。从多个角度分别对两组数据设计用户表征方式,并详细阐述了各种表征下特征的含义。在给定的用户基本属性与用户表征方式下,尝试了多种用户分类方法,进行了丰富的对比实验。3.对比了不同手机应用数据的用户属性挖掘能力。
[Abstract]:People are eager to get information and services from the Internet anytime and anywhere, which accelerates the development of mobile Internet, and promotes the mobile Internet to enter the era of "big data" for a long time. Smart phone has always been the main body of mobile Internet, in order to adapt to the explosive needs of users, technology innovation. Users frequently use smart phones. It produces a huge amount of personalized log data. This data is closely related to the behavior of the user, so it implies a very wide range of user attributes, such as gender. Personal information such as age as well as users' interests hobbies and living habits etc. Through excavating the user's various attributes we can deepen the comprehensive and objective understanding of the user and then help us to continuously improve the equipment. Services and applications. In recent years, more and more people begin to use smart phone related data to mine user attributes. Most of them are based on location information such as global positioning system, base station, wireless network hot spots and so on. The application of smart phone is the main entrance for users to use various services of mobile phone. The data related to mobile phone application can also reflect the basic attributes of users. Interest preferences, lifestyle habits, etc., provide us with a new opportunity to better understand the user. The list of apps installed on a smartphone and the sequence of data used by the app are very easy to collect. And contains a wealth of information. This paper specifically mining user attributes from these mobile phone application related data, completed the following work: 1. Put forward a user attribute mining scheme based on mobile phone application data. And around the application installation list data and the application uses the sequence data to design the user attribute mining concrete frame. 2. Elaborated the comprehensive data preprocessing process. Two groups of data are analyzed in a comprehensive way. Two groups of data are designed to represent the user from multiple angles, and the meaning of the features under various representations is elaborated in detail. Under the given user basic attributes and user representation. This paper attempts a variety of user classification methods and carries out rich comparative experiments. 3. The user attribute mining ability of different mobile phone application data is compared.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【相似文献】