Filling missing values for AI-based (load) forecasts within the InterFlex micro grid demo in Simris, Sweden

dc.contributor.affiliationRWTH Aachen
dc.contributor.affiliationRWTH Aachen
dc.contributor.affiliationEON Energy Research Center - RWTH Aachen
dc.contributor.affiliationE.ON
dc.contributor.affiliationEON Energy Research Center - RWTH Aachen
dc.contributor.authorPohlmann, Roxana
dc.contributor.authorWilms, Henning
dc.contributor.authorCupelli, Marco
dc.contributor.authorElgezua Fernandez, Inko
dc.contributor.authorMonti, Antonello
dc.contributor.countryGermany
dc.contributor.countryGermany
dc.contributor.countryGermany
dc.contributor.countryGermany
dc.contributor.countryGermany
dc.contributor.detailedauthorPohlmann, Roxana, RWTH Aachen, Germany
dc.contributor.detailedauthorWilms, Henning, RWTH Aachen, Germany
dc.contributor.detailedauthorCupelli, Marco, EON Energy Research Center - RWTH Aachen, Germany
dc.contributor.detailedauthorElgezua Fernandez, Inko, E.ON, Germany
dc.contributor.detailedauthorMonti, Antonello, EON Energy Research Center - RWTH Aachen, Germany
dc.date.accessioned2019-07-24T12:44:44Z
dc.date.available2019-07-24T12:44:44Z
dc.date.conferencedate3-6 June 2019
dc.date.issued2019-06-03
dc.description.abstract                                                Missing data impairs the performance of most neural networks with a particularly strong effect on time series prediction networks. Imputation addresses this issue and by replacing missing values with substitute values. The choice of a suitable imputation method requires fundamental knowledge of the dataset.                   Autoencoders (AE) have been widely applied in representation learning and feature extraction. In this paper we use a stacked denoising overcomplete autoencoder for imputation in multi-variate time series. We assess the model’s feature reproduction capability and compare its effect to simple mean imputation on a open source data set. Moreover, we assess the imputation’s influence on a recurrent neural network’s short-term load forecasting results and show that our proposed autoencoder model yields better results in feature imputation and significantly improves the forecasting accuracy for low and high fractions of missing data.
dc.description.conferencelocationMadrid, Spain
dc.description.conferencenameCIRED 2019
dc.description.openaccessYes
dc.description.peerreviewedYes
dc.description.sessionDistributed energy resources and efficient utilisation of electricity
dc.description.sessionidSession 4
dc.identifier.isbn978-2-9602415-0-1
dc.identifier.issn2032-9644
dc.identifier.urihttps://cired-repository.org/handle/20.500.12455/593
dc.identifier.urihttp://dx.doi.org/10.34890/817
dc.language.isoen
dc.publisherAIM
dc.relation.ispartProc. of the 25th International Conference on Electricity Distribution (CIRED 2019)
dc.relation.ispartofseriesCIRED Conference Proceedings
dc.titleFilling missing values for AI-based (load) forecasts within the InterFlex micro grid demo in Simris, Sweden
dc.title.number184
dc.typeConference Proceedings
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
CIRED 2019 - 184.pdf
Size:
375.29 KB
Format:
Adobe Portable Document Format