Predicting mobile apps spread: An epidemiological random network modeling approach

This web page contains the experiments and additional material for the paper J.Alegre-Sanahuja, J.-C. Cortés, F.-J. Santonja and R.-J. Villanueva, Predicting mobile apps spread: An epidemiological random network modeling approach, submitted to be considered for its publication in the journal SIMULATION: Transactions of The Society for Modeling and Simulation International.

The main objective of this study is to build a random network model and use them to predict mobile apps spread. The mobile applications business is a really big market growing constantly and, in app marketing, a key issue is to predict future app installations. The influence of the peers seems to be very relevant when downloading apps. Therefore, the study of the evolution of mobile apps spread may be approached using a proper network model that considers the influence of peers. Influence of peers and other social contagions have been successfully described using models of epidemiological type. Hence, in this study we propose an epidemiological random network model with realistic parameters to predict the evolution of downloads of apps. With this model, we are able to predict the behavior of an app in the market in the short term looking at its evolution in the early days of its launch. The numerical results provided by the proposed network are compared with data from real apps. This comparison shows that predictions improve as the model is feedback. Marketing researchers and strategy business managers can benefit from the proposed model since it can be helpful to predict app behavior over the time anticipating the spread of an app.