Abstract: Internet Detection System (IDS) can be used to detect a malicious attempt on the network or system aims to access restricted information or exploit internal resources, monitors and analyze user activity and maintain data integrity. Based on its detection mechanism, IDS can be grouped into offline and real-time IDS. On offline IDS, a saved labeled data set as in KDD Cup99 data set are used to measure the fitness factor of the rules on the identifier and we can analyze it to prevent some attacks happen in the future. Some classification methods have been widely used to classify IDS data set. It was used to recognize the pattern of the attacks so that we can differ between normal and unusual behaviors. On this research, we use Possibillistic C-Means (PCM) method as a classifier for KDD Cup99 data set. Based on the experiment, the best classification results was reach on 13% training data set with accuracy 68,63%. The accuracy is still low since PCM use several values of parameters and it affects the algorithm performances when the chosen values are not the best ones.
Aini Suri Talita and Eri Prasetyo Wibowo, 2020. Intrusion Detection Systems Data Classification by Possibilistic C-Means Method. Journal of Engineering and Applied Sciences, 15: 1170-1174.