AI and Machine Learning for Industrial Security With Level Discovery Method

Abstract

Protecting enterprise information security is a main task of Internet of Things system. The interaction between employees in same enterprise is based on level structures. So it is important to discover levels of employees for urban developers to protect enterprise information security. In this article, we propose a level discovery method for employees (LDME) from the records of employees using mobile phones named LDME. The call behavior between employees are expressed as several weighted directed complex networks, LDME represent edges in these weighted directed complex networks as vectors to exact both direction and weight information of the edges. Combined with supervised learning method, LDME prune these weighted directed networks into directed acyclic networks, which accurately reflect levels information between employees. At the same time, LDME mines the maximal frequent directed acyclic substructure from the above directed acyclic networks with efficient way, which indicate the stable levels information. We use real data to verify the performance of our method. The experimental result shows that the level of employees mined with our method is accurate and stable.