Conceptualization of intelligent clustering methodology for terrorists acts classification in nigeria

This paper aims at conceptualizing a platform for classifying terrorist acts using neuro-fuzzy and clustering techniques. The main objective is to design a knowledge base (comprising of knowledge warehouse, neural network (NN), fuzzy logic) and a neuro-fuzzy clustering system for Terrorists acts classification and also to propose a model for assigning numerical weights to qualitative Terrorists attributes. The conceptualization of neuro-fuzzy and clustering technique for classification of terrorists’ acts is presented. The conceptualization involves the design of Knowledge Base which is an integration of Knowledge Warehouse, Fuzzy Logic and Neural Network (NN). The Knowledge Warehouse is an abstraction of intelligent information which provides the decision maker with an intelligent analysis platform that improves all phases of the knowledge management process. The star and the snowflake schemas are used as the building blocks. The star schema consists of a large central table known as fact table and a set of smaller attendant tables known as dimension tables displayed in a radial pattern around the fact table. The snowflake schema consists of a central fact table and a set of constituent dimension tables which can be further broken down into sub-dimension tables. The fuzzy logic component of the system consists of fuzzification, fuzzy inference engine and defuzzification. The fuzzy logic component provides the inference under cognitive uncertainty while the neural network (NN) component offers adaptation, parallelism, fault tolerance and generalization in the system. The Neural Network (NN) is a three layer feed forward Neural Network (NN) architecture with nine nodes at the input layer, six nodes at the hidden layer, one node at the output layer representing the fatality measure of the terrorists acts generated through the connection weights of the hidden layer using sigmoid transfer function. The computed output is compared with the desired output and the difference, which is the error term, is used to adjust the connection weights by means of a back-propagation algorithm. The process is repeated until the error term is within the acceptable threshold. The inference engine comprises the Fuzzy C-Means (FCM) Clustering and Adaptive Neuro-Fuzzy Inference System (ANFIS)