Inter-event analysis and characterization of patterns in email communication
Abstract
The inter-event distribution of sent and received messages of each user in the Enron dataset was analyzed in this paper. Patterns were observed in the inter-event distributions. It was found out that one of the peaks formed, with value of Δtr and Δts equal to 64 hours, can be explained by repetitive weekly human activity. We also characterized the presence of burst in email inter-event for received and sent messages at values of Δt approximately equal to integral multiples of 60 seconds. The bursts at the specified Δt follow a power-law relation y=kΔt-a. The power-law exponents were a ≈ 0.6 and a ≈ 1 for received and sent messages, respectively.