The combination of neural estimates in prediction and decision problems

In this dissertation, different ways of combining neural predictive models or neural-based forecasts are discussed. The proposed approaches consider mostly Gaussian radial basis function networks, which can be efficiently identified and estimated through recursive/adaptive methods. Two different ways of combining are explored to get a final estimate – model mixing and model synthesis –, with the aim of obtaining improvements both in terms of efficiency and effectiveness. In the context of model mixing, the usual framework for linearly combining estimates from different models is extended, to deal with the case where the forecast errors from those models are correlated. In the context of model synthesis, and to address the problems raised by heavily nonstationary time series, we propose hybrid dynamic models for more advanced time series forecasting, composed of a dynamic trend regressive model (or, even, a dynamic harmonic regressive model), and a Gaussian radial basis function network. Additionally, using the model mixing procedure, two approaches for decision-making from forecasting models are discussed and compared: either inferring decisions from combined predictive estimates, or combining prescriptive solutions derived from different forecasting models. Finally, the application of some of the models and methods proposed previously is illustrated with two case studies, based on time series from finance and from tourism.

Short-term secular change in height, body mass and Tanner-Whitehouse 3 skeletal maturity of Madeira youth, Portugal

Secular trends in height and weight are reasonably well documented in Europe. Corresponding observations for skeletal maturation are lacking. Aim: To assess secular trends in height, body mass and skeletal maturity of Portuguese children and adolescents and to provide updated reference values for skeletal maturity scores (SMSs). Subjects and methods: Data for 2856 children and adolescents of 4–17 years, 1412 boys and 1444 girls, from The ‘Madeira Growth Study’ (MGS; 1996–1998) and from the‘Healthy Growth of Madeira Children Study’ (CRES; 2006) were used. Height and body mass were measured. Skeletal maturity was assessed with the Tanner-Whitehouse 2 and 3 methods. Results: Children from CRES were taller and heavier than peers from MGS. Differences in height reached 5.8cm in boys and 5.5cm in girls. RUS SMSs did not differ consistently between surveys boys, while higher RUS scores were observed in CRES girls. Adult RUS SMSs for MGS and CRES combined were attained at 15.8 years in boys and 14.8 years in girls. Corresponding ages for adult Carpal SMSs were 14.4 and 14.0, respectively. Conclusion: The short-term trends for height and mass were not entirely consistent with the trends in RUS and Carpal SMSs and SAs.