Electricity price forecasting &amp;#x2014; ARIMA model approach

Jakaša, Tina; Andročec, Ivan; Sprcic, Petar

doi:10.1109/eem.2011.5953012

Cited by 54 publications

(40 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…sharp peaks and dips). On the other hand, there are more successful results in studies like those of Contreras et al [4], Conejo et al [1], Weron and Misiorek [9], and Jakaša et al [37] since lags of shorter than 24 are used. Similarly, if the first and second differences are included in our model as in Eq.…”

Section: Arima Modelingmentioning

confidence: 95%

Forecasting the day-ahead price in electricity balancing and settlement market of Turkey by using artificial neural networks

Kölmek¹,

Navruz²

2015

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

Abstract:In determination of electric energy price, most price information coming from bilateral contracts is effective, but the importance of the spot market (pool market) price cannot be ignored. Forecasting the spot market price is very crucial, especially for companies actively participating in the spot market and giving purchase and sale bids. An artificial neural network is a way frequently used for price forecasting research. In this study, simulation studies about price modeling via artificial neural networks and proper artificial neural network configurations are examined. After selection of different network topologies and parameters, attempts are made to observe network performance by error rates and find the best suitable configuration. Moreover, a time series model is made and it is compared with the artificial neural network's error performance.

show abstract

Section: Arima Modelingmentioning

confidence: 95%

Forecasting the day-ahead price in electricity balancing and settlement market of Turkey by using artificial neural networks

Kölmek¹,

Navruz²

2015

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

show abstract

“…However, these similar approaches tend to deal with macroeconomic predictions rather than small size homes' or offices' loads as our case study. The investigations which focus their predictions on the energy field, as is our case, aim to price forecasting [35], and alternatively, to energy load calculation [36], which is also a correlated measure to also predict energy prices. These studies discuss the appropriateness of applying ARIMA models versus other approaches such as Artificial Neural Networks (ANN) or regressions models, to the specific problematic that they face.…”

Section: Related Workmentioning

confidence: 99%

Reducing Energy Waste through Eco-Aware Everyday Things

López-de-Armentia

Casado-Mansilla

López-Pérez

et al. 2014

Mobile Information Systems

View full text Add to dashboard Cite

Abstract. Society wastes much more energy than it should. This produces tons of unnecessary CO2 emissions. This is partly due to the inadequate use of electrical devices given the intangible and invisible nature of energy. This misuse of devices and energy unawareness is particularly relevant in public spaces (offices, schools, hospitals and so on), where people use electrical appliances, but they do not directly pay the invoice to energy providers. Embedding intelligence within public, shared appliances, transforming them into Eco-aware things, is valuable to reduce a proportion of the unnecessarily consumed energy. To this end, we present a twofold approach for better energy efficiency in public spaces: 1) informing persuasively to concerned users about the misuse of electronic appliances; 2) Customizing the operating mode of this everyday electrical appliances as a function of their real usage pattern. To back this approach, a capsule-based coffee machine placed in a research laboratory has been augmented. This device is able to continuously collect its usage pattern to offer feedback to coffee consumers about the energy wasting and also, to intelligently adapt its operation to reduce wasted energy. To this aim, several machine learning approaches are compared and evaluated to forecast the next-day device usage.

show abstract

“…ARIMA uses historical time series patterns and therefore, does not require the dependent variable; instead, time series information is used to generate the series itself. ARIMA is relatively simple and has been applied to forecast electricity loads and prices in previous studies [17][18][19][20]. Therefore, we explain the core of the ARIMA model here and further information can be found in [21,22].…”

Section: Arima Modelmentioning

confidence: 99%