A Case Study on Information Fusion Modelling in Email Archives

Information fusion modeling is quickly becoming a pioneer for disseminating and accumulating intelligence in the field of development and analysis of technological and scientific research. It also has an enormous impact on the study of social networks and business analysis. Due to the evolution of Internet technology, the flow of information has greatly increased in the contemporary life of people. The paper sets out an information fusion modeling study based on an email archive for the analysis of certain communication networks. The proposed study is on the borderline of graph theory and machine learning. It is suitable for big data with timestamps and it offers a unique way of their exploration also in an unsupervised manner. The proposed framework uses only the flow control from the email metadata set for the fusion model. The descriptive statistics approach was applied to evaluate the dataset’s quantitative properties such as frequency count of total mail sent and received, percentage coverage, and the intersection of sender and receiver. Furthermore, the fusion model mined for topological features in these temporal graphs and analyzed the change in the communication pattern over time.