Temporal Pattern of (Re)tweets Reveal Cascade Migration

Abstract

Twitter has recently become one of the most popular online social networking websites where users can share news and ideas through messages in the form of tweets. As a tweet gets retweeted from user to user, large cascades of information diffusion are formed over the global network. Existing works on cascades have mainly focused on predicting their popularity in terms of size. In this paper, we leverage on the temporal pattern of retweets to model the diffusion dynamics of a cascade. Notably, retweet cascades provide two complementary information: (a) inter-retweet time intervals of retweets, and (b) diffusion of cascade over the underlying follower network. Using datasets from Twitter, we identify two types of cascades based on presence or absence of early peaks in their sequence of inter-retweet intervals. We identify multiple diffusion localities associated with a cascade as it propagates over the network. Our studies reveal the transition of a cascade to a new locality facilitated by pivotal users that are highly cascade dependent following saturation of current locality. We propose an analytical model to show co-occurrence of first peaks with cascade migration to a new locality as well as predict locality saturation from interretweet intervals. Finally, we validate these claims from empirical data showing co-occurrence of first peaks and migration with good accuracy; we obtain even better accuracy for successfully classifying saturated and non-saturated diffusion localities from inter-retweet intervals.

Speeding up non-Markovian first-passage percolation with a few extra edges

“…k , the spreading curve ( T k , k/n) exhibits 'uncontrolled large plateaux' which do not decrease and do not converge as we increase the number of runs; see the dashed curves in Figure 1(a). Plateaux of similar type were also empirically found by Bhowmick [19]. On the other hand, adding just one extra edge to the tree (which does not change the local statistics of the graph) makes the spreading curve quite smooth; see the solid curves in Figure 1(a), and also Figure 1(b) where we present the SI spreading on the cycle with power-law exponent α ∈ ( 1 2 , 1).…”

“…Real-life examples with such ξ include, e.g. information cascades in Twitter™; see [19] and [24]. A striking feature of Figure 1: Spreading curves of the SI model simulation with power-law weights ξ with α = 0.8.…”

Temporal Sequence of Retweets Help to Detect Influential Nodes in Social Networks

“…Distinct temporal patterns present in the inter-retweet intervals of a cascade have been exploited to distinguish different user categories [35] as well as to identify different categories of retweeting activity on Twitter [36]. In addition, Bhowmick et al [37] have leveraged on inter-retweet intervals to detect cascade diffusion across multiple localities of the network. However, all these works have overlooked the potential of using pattern of inter-retweet intervals in a cascade to identify users playing key roles in diffusing information to different parts of the network; such users may serve as an indicator for identifying the influential nodes of a network.…”

“…Close inspection of T C reveals that there exists a peak in T C for some of the cascades, designating a significantly large inter-retweet time interval between two consecutive retweet events. Depending on the phase of occurrence of the first peak in T C , we classify cascades into the following two types [37] (see Fig. 1…”

“…However, as the locality gets saturated with the redundant tweet content, the post rarely gets retweeted in that locality, which in turn increases the inter-retweet intervals. Hence, the saturation in the diffusion locality [37] is reflected by the surge in inter-retweet time intervals. In this scenario, a retweet caused by the anchor node may diffuse the post to a new diffusion locality ( Fig.…”