Profiling Internet Users’ Activities using Fuzzy C-Means Algorithm

Abstract

The internet's openness has many benefits for consumers, like quick access, ease of use, and affordable software, but it also has drawbacks, such as being vulnerable to numerous cyberattacks that might harm businesses, governments, and individual individuals. Therefore, by looking at log files that display the pattern of users' actions in the network, profiling internet users using the Fuzzy C-Means (FCM) method is intended to give network managers a fast snapshot of the internet user's behavior within their managed network. Other attributes were applied to the data along with the FCM algorithm. MATLAB codes were used to implement the developed technique. The conventional FCM algorithm was used to assess the built system's performance in terms of the following metrics: sensitivity, precision, accuracy, specificity, and execution time. The investigation showed that the FCM algorithm performed better than the Simple K-Means (SKM) technique. Five (5) metrics were employed for the evaluation of the performance of the algorithms. The standard FCM showed slight improvement over the SKM. It was discovered that the standard FCM performed slightly better than the SKM with respect to all the metrics used in the evaluation apart from the execution time which can be trade off, since the system is intended to be used for control purpose and at such, accuracy is paramount. Hence, the high computational time of the standard FCM could be trade-off for better and accurate control system.