Forecasting the Monthly Electricity Demand of Georgia using Competitive Models and Advices for the Strategic Planning

Demir, Ahmet; Özsoy, Salih

doi:10.6007/ijarems/v3-i5/1189

Cited by 3 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…And no matter how accurate the ANFIS and LS SVM models look, they could not prove themselves to be the optimal model of choice for 4 and 9 out of the 12 months of the year respectively. These obvious shortcomings could be very costly in critical planning and decision-making processes in the power supply sector or the energy market [2]. Similar extensive comparisons have been made in various studies including that done and referenced in [12].…”

Section: Resultsmentioning

confidence: 68%

“…The amount of electricity to power a city, country or any sizeable load condition is usually not a storable commodity [2], therefore, power sector managers need to periodically make difficult but vital decisions regarding economic load dispatch, unit commitment, fuel allocation, off-line network analysis, purchase or generation of electric power, load switching, infrastructure development etc. [3].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

The Design and Demonstration of an Error Compensated Quasi-Hybrid Load Forecasting Technique

2023

IRJMETS

View full text Add to dashboard Cite

In the spirit of the evergreen drive to pool the respective merits of two or more forecast techniques so as to contrive a hybrid that is much more effective than any of its parents, this study x-rays the electricity demand forecast performance of three differently developed models, in order to extract and analyze their respective forecast errors in different months of the year, and then discover within the historical data, the models' respective epochs of comparative advantage. This information is then used to design a multi-model forecasting guide with error compensatory facility. Results show that none of the developed standalone models is useless in all facets of the historical data and none displays absolute superiority over others either; implying that each of the developed forecast model could be dedicated to use, only in that (or those) future period(s) in which each has been empirically proven to be most effective, year on year. The chart developed therefrom serves as a proposition to enhancing the forecast accuracy for researchers and industry players, both in the power sector as well as in other fields whose trend of historical data exhibit periodicity or seasonality similar to that in the power supply sector. This is so because standard metrics in this study tend to distinguish this forecast approach from its component models in the light of its obvious superior performance; being a hybrid forecasting technique of sorts.

show abstract

Section: Resultsmentioning

confidence: 68%

Section: Introductionmentioning

confidence: 99%

The Design and Demonstration of an Error Compensated Quasi-Hybrid Load Forecasting Technique

2023

IRJMETS

View full text Add to dashboard Cite

show abstract

“…Demir and Ozsoy (2014), have stated that forecasting is a complicated process as the factors such as innovation in technology, changes in culture and social values, unstable economic conditions, new product, stronger competitors, improved services, etc. There are different models for forecasting and their accuracies are depending on the situations and data considered.…”

Section: Previous Workmentioning

confidence: 99%

Modeling Trend in Telecommunication in Sri Lanka: A Case study on Internet and Cellular Connections

Karunarathna¹,

Brindha²,

Paramadevan³

2017

IJSP

View full text Add to dashboard Cite

The telecommunication is one of the modes of communication, in which most investments are made. It consists of internet, mobile phones, wired and wireless fixed phones, fax, televisions, radio and some other. Among them, demand for internet and cellular phones rapidly increases. For a smooth function of this business, knowledge on demand is much important. Effective forecasts help a business to manage its supply efficiently. This study aimed to find out an accurate mechanism for prediction of demand for internet conections and cellualr phone collections.Based on the secondary data available in central bank reports from 1996 to 2016, several statistical forecasting models were evaluated for an accurate prediction. There can be seen an increasing demand for both internet and cellular phone connections. Number of internet connections has gone up from 4 110 to 4 921 000, while the usage of cellular phones has developed from 71 228 to 26 228 000 during this period. Rapid growth in internet usage has happened after 2009, while after year 2003, usage of cellular phone has increased rapidly. With compared to models fitted for original form of data, models for log transformed data show better performances. The best performance in prediction of internet connection was given by ARIMA (1,1,1) model fitted for log transformed data, meanwhile ARIMA (0,1,2) model fitted for log transformed data showed the best fit for series of cellular connections. Double exponential smoothing models also show better fit for both series.

show abstract

Designing a Forecast Model for Economic Growth of Japan Using Competitive (Hybrid ANN vs Multiple Regression) Models

Demi̇r

Shadmanov²,

Aydınlı³

et al. 2015

Int J Econ Manag Sci

View full text Add to dashboard Cite

Forecasting the Monthly Electricity Demand of Georgia using Competitive Models and Advices for the Strategic Planning

Cited by 3 publications

References 26 publications

The Design and Demonstration of an Error Compensated Quasi-Hybrid Load Forecasting Technique

The Design and Demonstration of an Error Compensated Quasi-Hybrid Load Forecasting Technique

Modeling Trend in Telecommunication in Sri Lanka: A Case study on Internet and Cellular Connections

Designing a Forecast Model for Economic Growth of Japan Using Competitive (Hybrid ANN vs Multiple Regression) Models

Contact Info

Product

Resources

About