Viral marketing branching processes

We consider the inherent timeline structure of appearance content in online social networks (OSNs) while studying propagation. model propagation a post/content interest by an appropriate multi-type branching process. The process allows one to predict emergence global macro properties (e.g., spread post network) from laws and parameters that determine local interactions. interactions largely depend upon (an inverse stack capable holding many posts dedicated each user) number friends (i.e., connections) users, etc. explore use processes analyze viral post, e.g., derive expected shares, probability virality content, In OSNs, new push down existing contents timelines, which can greatly influence propagation; our analysis considers this influence. find leads draw incorrect conclusions when (TL) is ignored: (a) for instance, even less attractive are shown get viral; (b) ignoring TL also indicates erroneous growth rates. More importantly, cannot capture some interesting paradigm shifts/phase transitions; example, chances not monotone with network activity parameter, as including last part, we integrate auctions into marketing model. study optimization problem considering real-time bidding. again compared without varying levels network. fails relevant phase transitions, thereby making incomplete.