Synthesis of multi-year PV production data using generative adversarial networks

Kimball, Gregory M.; Pauchet, Camille M.; Ghadami, Rasoul; Zaragoza, Alberto Fonts

doi:10.1109/pvsc43889.2021.9518979

Search citation statements

Order By: Relevance

Paper Sections

Select...

Gan -Based Methods1

Citation Types

Supporting

Mentioning

Contrasting

Year Published

2021

2023

Publication Types

Select...

Preprint1

Other1

Relationship

Self Cite0

Independent2

Authors

Journals

Cited by 2 publications

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MedGAN [53] was the first GAN to model high dimensional, multi-label discrete variables in electronic health records (EHRs). GANs have been used to generate time-series data from many domains such as physiological signals [54], medical ICU data [16], financial time-series [19], [18], and PV (photovoltaic) production [55]. The TimeGAN [56] and DoppelGANger [57] are frameworks that modify the traditional GAN architecture to make it more suitable for time-series data.…”

Section: Gan -Based Methodsmentioning

confidence: 99%

Is Synthetic The New Real? Performance Analysis of Time Series Generation Techniques with Focus on Network Load Forecasting

Naveed¹,

Hashmi²,

Tajved³

et al. 2021

Preprint

View full text Add to dashboard Cite

This paper explores whether Generative Adversarial Networks (GANs) can produce realistic network load data that can be utilized to train machine learning models in lieu of real data. In this regard, we evaluate the performance of three recent GAN architectures on the Telecom Italia data set across a set of qualitative and quantitative metrics. Our results show that GAN generated synthetic data is indeed similar to real data and forecasting models trained on this data achieve similar performance to those trained on real data.

show abstract