Lidé

Ing. Pavel Šedek

Všechny publikace

THE PREDICTION OF TROPOSPHERIC OZONE USING A RADIAL BASIS FUNCTION NETWORK

  • Autoři: Kříž, R., Ing. Pavel Šedek,
  • Publikace: ISCS 2014: Interdisciplinary Symposium on Complex Systems. Dordrecht: Springer, 2015, pp. 115-123. Emergence, Complexity and Computation. ISSN 2194-7287. ISBN 978-3-319-10758-5. Available from: http://link.springer.com/chapter/10.1007%2F978-3-319-10759-2_13
  • Rok: 2015
  • DOI: 10.1007/978-3-319-10759-2_13
  • Odkaz: https://doi.org/10.1007/978-3-319-10759-2_13
  • Pracoviště: Centrum znalostního managementu
  • Anotace:
    The goal of this paper is to analyze the tropospheric ozone (O3) concentration time series and its prediction using artificial neural networks (ANNs). Tropospheric ozone has harmful effects on human health and on the environment. This study was based on daily averaged tropospheric ozone (O3) data from Pardubice in the Czech Republic. In this study, daily averaged ozone concentrations in Pardubice were predicted using a radial basis function network (RBFN) with three pollutant parameters and three meteorological factors in selected areas. We used a three-layer ANN, which consists of input, hidden, and output layers.

RECURRENCE PLOTS OF ELECTRICITY PRICE SERIES AND THE PREDICTION OF ELECTRICITY PRICES USING ARTIFICIAL NEURAL NETWORKS

  • Autoři: Kříž, R., Ing. Pavel Šedek,
  • Publikace: POSTER 2014 - 18th International Student Conference on Electrical Engineering. Prague: Czech Technical University, 2014, ISBN 978-80-01-05499-4.
  • Rok: 2014
  • Pracoviště: Centrum znalostního managementu
  • Anotace:
    Electricity is one of those commodities with which price is highly volatile and with considerable number of jumps. That is especially caused by the fact that the electricity is not possible to store. Thus the prediction of electricity price is a really complicated problem. In this paper, we focus on analyzing the electricity prices using recurrence plots and predicting it using radial basis artificial neural networks.At first we will estimate the time delay and the embedding dimension, which is needed for recurrence plots. Later we focus on building model using artificial neural networks to predict future prices and analyze prediction horizon.

Electricity intra-day spot prices prediction via artificial neural networks

  • Autoři: Ing. Pavel Šedek,
  • Publikace: Smart Art. Praha: České vysoké učení technické v Praze, Fakulta elektrotechnická, 2013.
  • Rok: 2013
  • Pracoviště: Katedra ekonomiky, manažerství a humanitních věd
  • Anotace:
    The ability to predict future electricity prices on the spot market provides possibility for electrical devices to optimize their power consumption based on operating costs. This approach is an alternative to receiving control singals from the smart grid. It could be used when connected to an old-fashioned electrical grid without any infrastructure change.

Za stránku zodpovídá: Ing. Mgr. Radovan Suk