With the advancement of technology, a surge of research interest in cognitive radio based networks in smart communities has been mounting. It is anticipated that CR-enabled networks will play a vigorous role in the enrichment of communication efficiency in neighborhood sensor area network. This paper presents a framework for Cognitive Radio based event anomaly detection mechanism. A skeleton for intelligent learning, detection and decision mechanism for Local Controller Unit and a Primary Controller Unit is also proposed and discussed in the model. The proposed model has four distinct layers namely sensors, routers, Local Controller Unit and Primary Controller Unit. A scheme for emergency situation detection and notification has been proposed. This paper also introduces a cluster formation scheme for better accuracy in data transmission among different hierarchical layers. The network module of the proposed model is later simulated and validated for some important performance communication metrics.