Daniel Peña Sánchez de Rivera
Departamento de Estadística e Instituto de Big Data Financiero. Universidad Carlos III
Se analiza como la disponibilidad creciente de datos masivos (Big Data) está cambiando la forma en que aprendemos. Se estudian nuevos métodos estadísticos en siete campos: aprender de datos complejos, visualizarlos en muchas dimensiones, controlar los problemas de contrastes múltiples, considerar la heterogeneidad, seleccionar modelos y reglas de predicción de forma automática, estimar con datos masivos y utilizar la información de redes para mejorar los modelos de predicción. Se presentan ejemplos de estos campos en problemas reales donde se dispone de información masiva. Se concluye que aprender de datos masivos requiere integrar métodos estadísticos, de investigación operativa y de ciencias de la computación.
Ana Paula Barbosa-Póvoa
CEG-IST Instituto Superior Técnico
Universidade de Lisboa
Sustainable Supply Chains (SSC) are complex network systems of entities that manage products from suppliers to customers and associated returns while considering simultaneously social, environmental and economic objectives. The management of such systems has recently gained a significant importance as companies face the task of incorporating sustainable concerns in their activities caused by growing society awareness towards environmental and societal problems. But being per se supply chains complex systems the enlargement of their common goal, profit, towards sustainability goals results in very challenging and difficult problems. To solve such problems there is a need of using decision tools to support decisions makers and in these Operational Research (OR) methods are fundamental. In this talk we explore the use of OR methods, more precisely optimization, to address the design, planning and operation of sustainable supply chains where a solution of compromise between economic, environmental and social objectives is trailed. Perspectives in the area are discussed and important challenges identified, which may drive research in sustainable supply chains when OR methods are to be explored.
University of Southampton
Over the past decades reusing data originated from administrative sources has gathered greater momentum internationally. The more recent digital revolution has greatly improved the availability of data outside the administrative sources, which can as well be repurposed for Official Statistics. In this talk I will summarise the changing paradigms of Official Statistics and the transformation of national statistical systems. Several high-profile on-going developments of administrative and unconventional big data sources will be reviewed and discussed. Special attention is given to situations where such sources can be envisaged to replace altogether the traditional sampling survey data, but the potential bias completely dominates the variance due to the large amount of data. An audit sampling approach will be developed and illustrated, where the aim is to statistically validate the estimation based on alternative sources, and to provide appropriate uncertainty measures.