Implementation of an intelligent clustering methodology for classification of terrorist acts

Terrorist acts have elevated the level of violence, intimidation and pose a threat to life/property, peace and security in the world today. Deployed solutions to curb the occurrence of terrorism prove to be of insignificant value, hence there is the need for more solutions. The research aims at implementing an intelligent clustering methodology for classification of the acts of terrorism in Nigeria. Three experiments were carried out. In the first experiment, the qualitative terrorists data attributes were converted to quantitative attributes using an existing One-of-N (OoN) method and the processed data supplied to Adaptive Neuro-Fuzzy Inference System (ANFIS) (OoN-ANFIS) for training. The second experiment converted the qualitative data attributes to quantitative attributes using the formulated Rank-Frequency-Based (RFB) model before the data was supplied to ANFIS (RFB-ANFIS) for training. In the third experiment, which constitutes the current study, the RFB-processed data was used by Fuzzy C Means (FCM) to generate initial membership values for each point in the data set and then supplied to ANFIS (RFB-FCMANFIS). The results show that RFB-FCMANFIS model generated the least Root Mean Square Error (RMSE), Mean Absolute Error (MAE), training error and checking error of 0.002887, 0.004598, 0.0000713 and 0.0056155 respectively with the highest correlation coefficient of 0.99954, therefore indicating a superior classification capability using the RFB-FCMANFIS.