Suchergebnisse

This paper presents an alternative calculation procedure to calculate the mortality rate, exploiting the data available in the Eurostat demography database for Spain. This methodology has been devised based on two of the most widely known and widespread models to establish the mortality rate: The Gompertz-Makeham (GM) and Lee-Carter (LC) models. Our main goal is to obtain a model yielding a similar accuracy than LC or GM, but able to capture the variation of their parameters over time and ages. The method proposed herewith works by applying simple or double fitting, with non-linear functions, to the values of the parameters considered by each one of such models. One of the main advantages of our approach is that we considerably reduce the amount of data that is required to establish the mortality rate, with respect to what would be needed if the traditional models were used. On the other hand, it also allows analyzing the evolution of the mortality rate, even if no real data was available for a particular year. The results evince that, besides fulfilling the two aforementioned goals, the proposed scheme yields an estimation error that is comparable with that offered by the traditional approach. ; This work has been supported in part by the Spanish Government (MINECO) and Fondo Europeo de Desarrollo Regional (FEDER) by means of the ADVICE (TEC2015-71329) project. Ramon Aguero was also partially supported by the Future Internet Enabled Resilient Cities (FIERCE) project, RTI2018-093475-A-100 (MCIU/AEI/FEDER/UE).

"The final publication is available at Springer via http://dx.doi.org/10.1007/s10588-013-9156-z ; We present a new mathematical model that predicts the number of users informed and influenced by messages that are propagated in an online social network. Our model is based on a new way of quantifying the tie-strength, which in turn considers the affinity and relevance between nodes. We could verify that the messages to inform and influence, as well as their importance, produce different propagation behaviors in an online social network. We carried out laboratory tests with our model and with the baseline models Linear Threshold and Independent Cascade, which are currently used in many scientific works. The results were evaluated by comparing them with empirical data. The tests show conclusively that the predictions of our model are notably more accurate and precise than the predictions of the baseline models. Our model can contribute to the development of models that maximize the propagation of messages; to predict the spread of viruses in computer networks, mobile telephony and online social networks. ; This work was carried out by the financing of the Ecuadorian government of President Rafael Correa D. and by FLAMINGO, a Network of Excellence project (318488) supported by the European Commission under the Seventh Framework Program and the project TEC2015-71329-C2-2-R (MINECO/FEDER) from Ministerio de Economía y Competitividad. R.M.O-G. thanks Dr. Joan Serrat Fernández, Dr. Xavier Muñoz and Dr. Josep Fàbrega, professors at the University Polytechnic of Catalonia. Similarly, Dr. Esteban Samaniego Alvarado and PhD student Vladimiro Tobar Solano, professors at Universidad de Cuenca, and the PhD student Lucía Mendez Tapia for the accomplishment of this work. ; Peer Reviewed ; Postprint (author's final draft)