Predicting a User’s Numeric Identity from the Search of Attribute Data

Abstract

In common Internet environments, most of the websites or services constrain the user account creation. Since the Internet is accessible by all and offers more and more services, a user has several accounts on the web. The difficulty in controlling their accounts does not leave indifferent to the users of the web. Hence the use of easy or insecure passwords. This is why we are victims of attacks and forgetting our passwords. Large companies such as Facebook, Google, etc., offer authorization and authentication mechanisms using the Oauth and OpenID protocol, which requires the opening of an account. To be independent of a social network or a site, it would be important to develop a model to make a statistical analysis between the attributes of the profiles of the same user and to create an account. Using the same password for all its different accounts could be an approach but avoiding the proliferation of data by proposing a model of identity analysis would be even more interesting. That is why this article proposes a centralized account management model by making a comparative and statistical study of the identity attributes and proposing a single account to the user to manage all its different accounts. So, we have a horizontal analysis between the attributes of the identity categories and a vertical analysis between these categories. This study allowed us to find a threshold to conclude that an account belongs to a user.