El 29 de junio tendrá lugar el Seminario: Predicción estadística en variables relacionadas con la producción de energía.
El 29 de junio a las 12:00 horas tendrá lugar en el Salón de Grados de la Facultad de Matemáticas de la Universidad de Santiago de Compostela (USC) el Seminario: Predicción estadística en variables relacionadas con la producción de energía.
El seminario se estructura en 2 charlas. La primera de ellas, impartida por Manuel Oviedo de la Fuente, Investigador contratado de ITMATI y titulada: “Selección de variables aplicada a la predicción de la demanda y el precio de la energía”. La segunda de ellas, será impartida por Jairo Cugliari, de la Universidad de Lyon 2, titulada “Non parametric forecasting and functional clustering using wavelets. Application to electricity demand”.
PROGRAMA:
- 12:00 – 13:00 Manuel Oviedo de la Fuente (Universidad de Santiago de Compostela e ITMATI) “Selección de variables aplicada a la predicción de la demanda y el precio de la energía”
Resumen:
Este trabajo considera el problema de la selección de variables cuando algunas de las variables tienen una naturaleza funcional (como la curva diaria del precio de la energía) que se relacionan con otro tipo de variables (escalares, multivariantes, direccionales, etc.). Nuestra propuesta parte de un modelo de regresión nulo y selecciona secuencialmente una nueva variable que se incorporará al modelo hasta que se incluya toda la información relevante y no redundante. Para ello se realiza un uso exhaustivo de la correlación de distancia, R, propuesta por Szekely et al. (2007) que permite cuantificar la dependencia entre dos variables de dimensiones arbitrariamente finitas. El procedimiento ha mostrado resultados prometedores cuando se aplica a conjuntos de datos reales. Como ilustración la propuesta se ha aplicado a la previsión de la demanda y precio de la energía en el mercado ibérico utilizando información procedente de diferentes fuentes como el tipo de energía generada (nuclear, carbón, ciclo combinado, renovables, etc.), datos meteorológicos (temperatura, radiación solar, velocidad del viento, etc.) y variables estacionales (año, mes, día de la semana, festivos, etc.).
- 13:00 – 14:00 Jairo Cugliari (Universidad de Lyon 2) “Non parametric forecasting and functional clustering using wavelets. Application to electricity demand”
Resumen:
This talk has an industrial motivation that is the nonparametric forecast of electricity demand for the French producer EDF. We then present two methods for detecting patterns and clusters in high dimensional time-dependent functional data. Our methods are based on wavelet-based similarity measures, since wavelets are well suited for identifying highly discriminant time-scale features. The multiresolution aspect of the wavelet transform provides a time-scale decomposition of the signals allowing to visualize and to cluster the functional data into homogeneous groups. For each input function, through its empirical orthogonal wavelet transform the first method uses the distribution of energy across scales to generate a representation that can be sufficient to make the signals well distinguishable. Our new similarity measure combined with a feature selection technique is then used within classical clustering algorithms to effectively differentiate among high dimensional populations. The second method uses similarity measures between the whole time-scale representations that are based on wavelet-coherence tools. The clustering is then performed using a k-centroid algorithm starting from these similarities. Finally the practical performance of these methods is illustrated through the daily profiles of the French electricity power demand involved in nonparametric forecasting as well as individual consumers clustering involved in the forecasting by disaggregation of the electricity consumption. The talk is related to joint works with Anestis Antoniadis, Xavier Brossat, Yannig Goude and Jean-Michel Poggi
This talk has an industrial motivation that is the nonparametric forecast of electricity demand for the French producer EDF. We then present two methods for detecting patterns and clusters in high dimensional time-dependent functional data. Our methods are based on wavelet-based similarity measures, since wavelets are well suited for identifying highly discriminant time-scale features. The multiresolution aspect of the wavelet transform provides a time-scale decomposition of the signals allowing to visualize and to cluster the functional data into homogeneous groups. For each input function, through its empirical orthogonal wavelet transform the first method uses the distribution of energy across scales to generate a representation that can be sufficient to make the signals well distinguishable. Our new similarity measure combined with a feature selection technique is then used within classical clustering algorithms to effectively differentiate among high dimensional populations. The second method uses similarity measures between the whole time-scale representations that are based on wavelet-coherence tools. The clustering is then performed using a k-centroid algorithm starting from these similarities. Finally the practical performance of these methods is illustrated through the daily profiles of the French electricity power demand involved in nonparametric forecasting as well as individual consumers clustering involved in the forecasting by disaggregation of the electricity consumption. The talk is related to joint works with Anestis Antoniadis, Xavier Brossat, Yannig Goude and Jean-Michel Poggi
INSCRIPCIÓN: La inscripción es gratuita: Formulario de inscripción (pinche aquí)
Se entregará un certificado de asistencia para aquellos que lo soliciten.
Colabora la Consellería de Cultura, Educación e Ordenación Universitaria de la Xunta de Galicia a través de la Red Tecnológica de Matemática Industrial (Red TMATI) y del convenio que ITMATI tiene con esta Consellería.
Fecha:
Jue, 2017-06-15 
