A Method of Identifying User Identity Based on Username Features

1. Introduction

With the rapid development of the internet, the scale of internet users is also growing, as of December 2016, the global total number of Internet users reached 3.5 billion people, accounting for the world's total population of 46.1%. At the same time, network applications also penetrate into all aspects of life, including the rapid growth in the number of registered users. For example, as of the third quarter of 2016, the microblogging service averaged at 317 million monthly active Twitter users, while for the well-known foreign website, Facebook, had 1.79 billion monthly active users. On the internet, people are usually registered to use different network applications for information access, chat communication and other activities. Research shows that (Bartunoy et al., 2012), about 8 4% of internet users have more than one social networking site account. As a result, there is a large number of duplicate user identities in the Internet. Integration of these duplicate user identity has gradually become a very meaningful research topic.

From a commercial point of view, user's repeated information in multiple network applications can be used to determine and integrate, to help network service providers a comprehensive understanding of network users, master user identity characteristics, thus providing better personalized service.

From the network security point of view, it can help the network security manager to discover the false or illegitimate identity by discriminating the repeated users 'information, so as to reduce the various network security risks caused by user anonymity and protect the legitimate user’s rights and interests. In addition, this study has a wide range of applications in collaborative recommendation, information retrieval and other fields.

However, due to the fact that single sign-on(SSO) technology is not widely used, and individual websites store their own user account information separately, so researchers have to find various ways to obtain user account information. In general, if user wasn’t register some site, they need to provide some personal information such as username, password, mailbox, age, address and so on. Intuitively, the user accounts with similar registration information in different network applications may belong to the same person. But for commercial interests or privacy protection needs, access to detailed and accurate registration information data is very difficult. In addition, for privacy and security point of view, the user registration provided is mostly incomplete or part of the content is false, which also adds to the difficulty of user authentication.

The user name usually consists of English letters, numbers and special characters such as “-”, “_” and other components, the length between 4 to 26. Compared with other registration information, username is easier to obtain, and will not reveal user privacy. In addition, the user name is usually able to reflect the owner's identity or personality characteristics, such as name, origin, birth date, preferences, etc.; and the same user's different user names usually have similar characteristics, such as length, character combination. Therefore, this paper defines the problem of user identity re-identification in multiple websites, and proposes a method to determine user identity based on user name characteristics.