## No full-text available

To read the full-text of this research,

you can request a copy directly from the authors.

This paper seeks to analyse and evaluate Sri Lanka's energy consumption forecasts, more specifically, electricity, petroleum, coal and renewable electricity consumption using a variety of parametric and nonparametric forecasting techniques. The Sri Lankan economy is emerging following the end of a prolonged civil war, and thus this topic is opportune as accurate forecasts of energy requirements are indispensable for sustaining the ongoing rapid economic expansion. We also consider evaluating the appropriateness of parametric and nonparametric energy forecasting methods for Sri Lanka. In addition, this paper marks the introductory application of SSA for forecasting in Sri Lanka, and the results from SSA are compared against the popular benchmarks of ARIMA, ETS, HW, TBATS, and NN. We find statistically significant evidence proving that the SSA model outperforms ETS and HW at forecasting renewable electricity consumption, and also, that the SSA model outperforms ARIMA and TBATS models at forecasting coal consumption. On average, the SSA model is found to be best for energy consumption forecasting in Sri Lanka whilst the Neural Networks model is second best.

To read the full-text of this research,

you can request a copy directly from the authors.

... In line with this, there were two methods based on frequency domain, which were Fourier Series Analysis (FSA) and Singular Spectrum Analysis (SSA). Both of them did not have to fulfill the parametric assumption, so that they were categorized in nonparametric methods[3][4]. Technically, SSA could divide data based on its elements such as trend, seasonal, cyclical, and noise. ...

This research aims to evaluate the performance of forecasting by Fourier Series Analysis (FSA) and Singular Spectrum Analysis (SSA) which are more explorative and not requiring parametric assumption. Those methods are applied to predicting the volume of motorcycle sales in Indonesia from January 2005 to December 2016 (monthly). Both models are suitable for seasonal and trend component data. Technically, FSA defines time domain as the result of trend and seasonal component in different frequencies which is difficult to identify in the time domain analysis. With the hidden period is 2,918 ≈ 3 and significant model order is 3, FSA model is used to predict testing data. Meanwhile, SSA has two main processes, decomposition and reconstruction. SSA decomposes the time series data into different components. The reconstruction process starts with grouping the decomposition result based on similarity period of each component in trajectory matrix. With the optimum of window length (L = 53) and grouping effect (r = 4), SSA predicting testing data. Forecasting accuracy evaluation is done based on Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). The result shows that in the next 12 month, SSA has MAPE = 13.54 percent, MAE = 61,168.43 and RMSE = 75,244.92 and FSA has MAPE = 28.19 percent, MAE = 119,718.43 and RMSE = 142,511.17. Therefore, to predict volume of motorcycle sales in the next period should use SSA method which has better performance based on its accuracy.

... In brief, the TBATS technique uses a new method that greatly reduces the computational burden in the maximum likelihood estimation when forecasting complex seasonal time series such as those with multiple seasonal peri- ods, high-frequency seasonality, non-integer seasonality, and dual-calendar effects (De Livera et al., 2011). TBATS has been used to forecast energy consumption (Silva & Rajapaksa, 2014), the price of gold (Hassani, Silva, Gupta, & Segnon, 2015) and housing downturns (Zietz & Traian, 2014) in previous studies. Thirdly, all the aforementioned techniques are classical meth- ods, and SSA is able to provide a completely different modelling approach as SSA is a filtering technique. ...

... The possible applications areas of Singular Spectrum Analysis are diverse: from mathematics and physics to economic and financial mathematics, from social science and market research to meteorology and oceanology (e.g. Hassani, 2007;Hassani, Mahmoudvand, & Zokaei, 2011;Hassani, Soofi, & Zhigljavsky, 2013a;Hassani & Zhigljavsky, 2008;Sanei, Ghodsi, & Hassani, 2010;Silva & Rajapaksa, 2014 and references therein). The aim of Singular Spectrum Analysis is to make a decomposition of the original series into the sum of a small number of independent and interpretable componentssuch as a slowly varying trend, oscillatory components and a structureless noise (Hassani, 2007). ...

The format of cycling time trials in England, Wales and Northern Ireland, involves riders competing individually over several fixed race distances of 10-100 miles in length and using time constrained formats of 12 and 24 h in duration. Drawing on data provided by the national governing body that covers the regions of England and Wales, an analysis of six male competition record progressions was undertaken to illustrate its progression. Future forecasts are then projected through use of the Singular Spectrum Analysis technique. This method has not been applied to sport-based time series data before. All six records have seen a progressive improvement and are non-linear in nature. Five records saw their highest level of record change during the 1950-1969 period. Whilst new record frequency generally has reduced since this period, the magnitude of performance improvement has generally increased. The Singular Spectrum Analysis technique successfully provided forecasted projections in the short to medium term with a high level of fit to the time series data.

... In this study, multiple linear regression models and ANN models are compared and one of the ANN models is chosen based on the model evaluation parameter. In 2014, Silva and Rajapaksa [20] applied a feed-forward neural network system with one hidden layer for energy consumption forecasting in Sri Lanka. They found that the ANN model outperforms ARIMA, ETS, Holt-Winters and TBATS models on average. ...

In the United States, the industrial sector is the driving engine of economic development, and the energy consumption in this sector may be considered as the fuel for this engine. In order to keep this sector sustainable (diverse and productive over time), energy plan-ning should be carried out comprehensively and precisely. In this study, an ANN model was applied to forecast the industrial energy demand and perform future projections for the period 2013–2030. Among all effective independent parameters on energy demand in the industrial sector, energy costs and GDP growth have been considered in this study based on correlation coefficient analysis. For the future trend of GDP, a second order polynomial equation is fitted to the GDP growth curve. For the other indepen-dent variables, we define three scenarios for potential future changes: Constant Price Scenario, Ascending Price Scenario, and Descending Price Scenario. The Constant Price and Descending Price scenarios show increases in energy demand, while results show that along with an increase in energy prices, the demand may decrease slightly. For comparison purposes, the results of the three scenarios are presented along with the predictions from the EIA presented in the Annual Energy Outlook 2013.

... In this paper we give consideration to both recurrent and vector forecasting algorithms as we evaluate the sensitiveness of SSA to outliers. In recent years, SSA has undergone various developments and continues to be increasingly applied to solve many practical problems (see, for example, [13][14][15][16][17][18][19][20][21][22][23][24]). ...

The aim of this paper is to study the effect of outliers on different parts of singular spectrum analysis (SSA) from both theoretical and practical points of view. The rank of the trajectory matrix, the magnitude of eigenvalues, reconstruction, and forecasting results are evaluated using simulated and real data sets. The performance of both recurrent and vector forecasting procedures are assessed in the presence of outliers. We find that the existence of outliers affect the rank of the matrix and increases the linear recurrent dimensions whilst also having a significant impact on SSA reconstruction and forecasting processes. There is also evidence to suggest that in the presence of outliers, the vector SSA forecasts are more robust in comparison to the recurrent SSA forecasts. These results indicate that the identification and removal of the outliers are mandatory to achieve optimal SSA decomposition and forecasting results.

... The roots of SSA are closely associated with King (1986a, 1986b). The applications of SSA are diverse and its growing success is evident in many different fields (see for example, Beneki & Silva, 2013;Ghil et al., 2002;Ghodsi, Hassani, Sanei, & Hicks, 2009;Hassani, Heravi, & Zhigljavsky, 2009;Hassani, Heravi, & Zhigljavsky, 2013;Hassani, Soofi, & Zhigljavsky, 2013;Hassani & Thomakos, 2010;Lisi & Medio, 1997;Silva, 2013;Silva & Rajapaksa, 2014). As noted above, there exist various different techniques which have been applied for forecasting tourism demand in the past. ...

Demand planning for electricity consumption is a key success factor for the development of any countries. However, this can only be achieved if the demand is forecasted accurately. In this research, different forecasting methods-autoregressive integrated moving average (ARIMA), artificial neural network (ANN) and multiple linear regression (MLR)-were utilized to formulate prediction models of the electricity demand in Thailand. The objective was to compare the performance of these three approaches and the empirical data used in this study was the historical data regarding the electricity demand (population, gross domestic product: GDP, stock index, revenue from exporting industrial products and electricity consumption) in Thailand from 1986 to 2010. The results showed that the ANN model reduced the mean absolute percentage error (MAPE) to 0.996%, while those of ARIMA and MLR were 2.80981 and 3.2604527%, respectively. Based on these error measures, the results indicated that the ANN approach outperformed the ARIMA and MLR methods in this scenario. However, the paired test indicated that there was no significant difference among these methods at α = 0.05. According to the principle of parsimony, the ARIMA and MLR models might be preferable to the ANN one because of their simple structure and competitive performance.

The forecasting of electricity consumption and demand plays a pivotal role in electric power systems planning. This paper proposes the combination of forecasts from two approaches with the aim of improving the forecasting accuracy, in order to make the best use of the installed transmission and generating capacity. In the first approach, the consumption time series is decomposed by wavelet analysis and a Box-Jenkins model is fitted to each wavelet component, following which the individual components forecasts are added to compute the total consumption forecast. The alternative approach, uses the Singular Spectrum Analysis technique to model the consumption time series in order to shrink the noise level. Thereafter, the Box-Jenkins model is used to forecast the filtered time series, producing a second forecast for the consumption series. Eventually, the two forecasts are combined geometrically in order to minimize the mean square error. The proposed methodology is illustrated by a computational experiment with the time series of residential consumption of electricity in Brazil.

The aim of this paper is to propose a new methodology for hydroelectric energy forecasting. A new approach for selection of the number of eigenvalues in SSA is also proposed. In this paper it is proposed the hierarchical clustering associated to PCA and integrated to ARIMA models. The proposed approach is applied to forecast the affluent flow in a hydroelectric plant located at Parana River Basin, Brazil. As a matter of fact, modeling such series is quite important for the optimal dispatch of the energy generation in Brazil due to the heavy participation of hydro plants in the country (over 85% of the generated energy comes from hydro plants).

The different forms of the multivariate singular spectrum analysis (SSA) and their associated forecasting algorithms are considered from both theoretical and practical points of view. The new multivariate vector forecasting algorithm is introduced and its uniqueness is evaluated. The performance of the new multivariate forecasting algorithm is assessed against the existent multivariate technique using various simulated and real data sets (namely European Electricity and Gas series). The forecasting results confirm that the performance of the new multivariate approach is more accurate than the current approach. The optimality of the window length and the number of eigenvalues in multivariate SSA are considered and various bounds are recommended. The effect of common components between two time series is evaluated through a simulation study. The concept of similarity and dissimilarity are also considered based on the matched components among series.

SUMMARY Univariate time series models make efficient use of available historical records of electricity consumption for short-term forecasting. However, the information (expectations) provided by electricity consumers in an energy-saving survey, even though qualitative, was considered to be particularly important, because the consumers' perception of the future may take into account the changing economic conditions. Our approach to forecasting electricity consumption combines historical data with expectations of the consumers in an optimal manner, using the technique of restricted forecasts. The same technique can be applied in some other forecasting situations in which additional information-besides the historical record of a variable-is available in the form of expectations.

The incessantly growing demand for energy consumption and the signi�cance of the
availability of sustainable energy for achieving long term economic growth de�nes the
importance of forecasting energy statistics. This paper analyses and forecasts actual
energy consumption data for EU-27 nations using both parametric and nonparametric
time series forecasting techniques. Singular Spectrum Analysis (SSA) is adopted as the
nonparametric time series analysis and forecasting technique and the results from SSA
are compared with ARIMA, which is a parametric forecasting technique.
Keywords: Electricity consumption; Renewable electricity consumption; Primary energy
consumption; Forecasting; Singular Spectrum Analysis; ARIMA; Parametric; Nonpara-
metric.

The demand for electricity in Sri Lanlca depends mainly on the activities ol'clomestic, industrial and comnlercial sectors ancl the three conlponents are highly correlatecl. Although such correlation does not affect univariate estimation procedures, it may lead to incorrect inferences of influerltial factors on the demand for electricity. As a result, separate univariate approacl~es for each sector may not bc an accur;~te niethocl of' identifying such factors. Therefore, this stucly alms to identify sucll facto1.s using multivariate regression whicb consjders the correlation among different sectors (or dependent variables) ancl estimates il multivariate demand model for the purpose offorecasting. The overall sign iricance of the fitted demand model and t he significant influential factors are assessed by multivariate tests such as Bartlett's using the statistical package SAS. Theoretically, demand is a function of its own price, the income level ,or consumers, and the price of substitutes. Gross Domestic Product (GDP) at constant (1960) factor prices is used as a proxy for income level of consumers and kerosene is tillten as a close substitute Ibl. elcctriclty. The analysis uses quarterly data Tot two periocls 1970-1977 ant1 1978-1994 to assess the effect of the liberalized economy int~~ocluccd in late 1977. During the period aftcr 1977, the effect ol'thc income levcl has i~lcreasecl substa~ltially clue to the liberalized economy. The substitution between electricity and lierosene is marginal ill the post-lil)ol~alizecl periocl, as electricity is more efficient antl- convenient tl~an Iterosene. Jlue to , such dit'ferenccs between the two pcriohs. the de~nancl for electricity lni>y be explained better by tnto models rather than a single lnoclel estimated for tl~e entire period. The multivariate demand model based on the post-li.beralized period is fbuncl to aclequately forecast the clemand for electricity. i

Interest in using artificial neural networks (ANNs) for forecasting has led to a tremendous surge in research activities in the past decade. While ANNs provide a great deal of promise, they also embody much uncertainty. Researchers to date are still not certain about the effect of key factors on forecasting performance of ANNs. This paper presents a state-of-the-art survey of ANN applications in forecasting. Our purpose is to provide (1) a synthesis of published research in this area, (2) insights on ANN modeling issues, and (3) the future research directions.

In Lebanon, electric power is becoming the main energy form relied upon in all economic sectors of the country. Also, the time series of electrical energy consumption in Lebanon is unique due to intermittent power outages and increasing demand. Given these facts, it is critical to model and forecast electrical energy consumption. The aim of this study is to investigate different univariate-modeling methodologies and try, at least, a one-step ahead forecast for monthly electric energy consumption in Lebanon. Three univariate models are used, namely, the autoregressive, the autoregressive integrated moving average (ARIMA) and a novel configuration combining an AR(1) with a highpass filter. The forecasting performance of each model is assessed using different measures. The AR(1)/highpass filter model yields the best forecast for this peculiar energy data.

An innovations state space modeling framework is introduced for forecasting complex seasonal time series such as those with multiple seasonal periods, high-frequency seasonality, non-integer seasonality, and dual-calendar effects. The new framework incorporates Box–Cox transformations, Fourier representations with time varying coefficients, and ARMA error correction. Likelihood evaluation and analytical expressions for point forecasts and interval predictions under the assumption of Gaussian errors are derived, leading to a simple, comprehensive approach to forecasting complex seasonal time series. A key feature of the framework is that it relies on a new method that greatly reduces the computational burden in the maximum likelihood estimation. The modeling framework is useful for a broad range of applications, its versatility being illustrated in three empirical studies. In addition, the proposed trigonometric formulation is presented as a means of decomposing complex seasonal time series, and it is shown that this decomposition leads to the identification and extraction of seasonal components which are otherwise not apparent in the time series plot itself.

Automatic forecasts of large numbers of univariate time series are often needed in business and other contexts. We describe two automatic forecasting algorithms that have been implemented in the forecast package for R. The first is based on innovations state space models that underly exponential smoothing methods. The second is a step-wise algorithm for forecasting with ARIMA models. The algorithms are applicable to both seasonal and non-seasonal data, and are compared and illustrated using four real time series. We also briey describe some of the other functionality available in the forecast package.

We consider the notion of qualitative information and the practicalities of extracting it from experimental data. Our approach, based on a theorem of Takens, draws on ideas from the generalized theory of information known as singular system analysis due to Bertero, Pike and co-workers. We illustrate our technique with numerical data from the chaotic regime of the Lorenz model.

The present study aims to forecast the primary energy demand in Turkey for the period 2007-2015using the Box-Jenkins methodology. The annual data for the period 1970-2006 provided by the Ministry ofEnergy and Natural Resources were used in the study. Considering the results of unit root test, energy demandseries is stationary at first difference. Later among alternative models it is found that the most appropriate modelis ARIMA (3,1,3) for energy demand series. According to this model, estimation findings show that the energydemand would continue its increasing trend also in the forecast period. It is expected that the primary energydemand will reach 119.472 TOE in 2015 with an approximately 22 percent increase compared to 2006.Therefore energy policies should be designed for increasing demand in Turkey.

We present and apply the Singular Spectrum Analysis (SSA), a relatively new, non-parametric and data-driven method used for signal extraction (trends, seasonal and business cycle components) and forecasting of the UK tourism income. Our results show that SSA outperforms slightly SARIMA and time-varying parameter State Space Models in terms of RMSE, MAE and MAPE forecasting criteria.

This paper develops a new approach for producing probabilistic wind power forecasts using a single forecast. The Singular Spectrum Analysis technique is used as the forecasting technique. Given the confidence interval calculated for the single forecast, a large number of random forecasts were generated through the Monte Carlo method. The purpose of generating probabilistic wind power forecasts is to use the results in a stochastic programming unit commitment problem. Therefore, probabilistic forecasts are reduced to a small number of representative forecast scenarios by applying a scenario reduction algorithm. The stability of the scenario reduction algorithm is also evaluated. The results indicate that the insignificant changes of operational cost of the stochastic programming problem reflect a deletion of unimportant forecast scenarios.

This paper seeks to provide an alternative forecast to that provided by the Energy Information Administration (EIA) on energy-related monthly CO2 emissions in the United States. The data on CO2 emissions from petroleum, natural gas, coal and total fossil fuels obtained via the EIA covering the period January 2005 to November 2013 is analysed and then forecasted using ARIMA, Holt-Winters, and Exponential Smoothing prior to introducing the Singular Spectrum Analysis (SSA) technique for CO2 emissions forecasting. A new combination forecast (EIA-SSA) is also introduced by merging the SSA and EIA forecasts, and is seen outperforming all models including the EIA forecast. Finally, the EIA-SSA model is used to provide an alternative 12 month ahead outlook for US energy-related CO2 emissions from December 2013 to November 2014. This research is expected to influence the methodology adopted by the EIA for forecasting CO2 emissions in the future by improving the accuracy of the forecasts, and the impact of this study will be clearer upon comparing the actual CO2 emissions in US with the EIA, and EIA-SSA forecasts over the 12 month period which follows.

Electricity demand forecasting could prove to be a useful policy tool for decision-makers; thus, accurate forecasting of electricity demand is valuable in allowing both power generators and consumers to make their plans. Although a seasonal ARIMA model is widely used in electricity demand analysis and is a high-precision approach for seasonal data forecasting, errors are unavoidable in the forecasting process. Consequently, a significant research goal is to further improve forecasting precision. To help people in the electricity sectors make more sensible decisions, this study proposes residual modification models to improve the precision of seasonal ARIMA for electricity demand forecasting. In this study, PSO optimal Fourier method, seasonal ARIMA model and combined models of PSO optimal Fourier method with seasonal ARIMA are applied in the Northwest electricity grid of China to correct the forecasting results of seasonal ARIMA. The modification models forecasting of the electricity demand appears to be more workable than that of the single seasonal ARIMA. The results indicate that the prediction accuracy of the three residual modification models is higher than the single seasonal ARIMA model and that the combined model is the most satisfactory of the three models.

The nature of electricity differs from that of other commodities since electricity is a non-storable good and there have been significant seasonal and diurnal variations of demand. Under such condition, precise forecasting of demand for electricity should be an integral part of the planning process as this enables the policy makers to provide directions on cost-effective investment and on scheduling the operation of the existing and new power plants so that the supply of electricity can be made adequate enough to meet the future demand and its variations. Official load forecasting in India done by Central Electricity Authority (CEA) is often criticized for being overestimated due to inferior techniques used for forecasting. This paper tries to evaluate monthly peak demand forecasting performance predicted by CEA using trend method and compare it with those predicted by Multiplicative Seasonal Autoregressive Integrated Moving Average (MSARIMA) model. It has been found that MSARIMA model outperforms CEA forecasts both in-sample static and out-of-sample dynamic forecast horizons in all five regional grids in India. For better load management and grid discipline, this study suggests employing sophisticated techniques like MSARIMA for peak load forecasting in India.

This paper traces the history of the electricity power sector in Sri Lanka from the introduction of electricity to the country by a private company called Baustead Brothers Ltd. in 1895 during the colonial period to date. The historical development provides an introduction to the present generating system of the country and serves also to highlight the principles of methodical development of power generation in the country. The main state sector agency in electricity generation and distribution is the Ceylon Electricity Board (CEB). The CEB is also entrusted with generation planning and it is empowered to make its decisions on the basis of least cost to the national economy. The country faced its most severe power crisis in 1996 and is facing another such crisis today. How CEB plans to meet the country's future electricity needs is thus discussed against this climate of power shortages and uncertainty. Providing all related information such as low growth in 1996 and economic scenarios that preceded these events the paper attempts to answer the key question, “Can we blame this on lack of planning and foresight on the part of CEB planners?” The CEB planning exercise is presented in detail and shows how the plans were developed in keeping in view the country's needs at various times. The generating options considered involve both conventional energy sources and renewables, and the selection is based on an analysis based in turn on the International Atomic Energy Agency's WASP 111 simulation model, the software of choice of CEB's generation planning division. It is shown that the Sri Lankan power system is in a state of transition from being predominantly hydro-based to predominantly coal-based within the first decade of this century. After giving details of why, how, and what, and outlining the capacity development plan after demand analysis, the paper concludes that the proof of the effectiveness of the plan will be in its implementation.

This paper describes a new model and a new generator of hourly wind speeds which were obtained using the Box-Jenkins method. All the steps leading to the determination of an autoregressive model are described. Tests were performed to verify the adequacy of the model and comparisons were made between generated and real series to check whether the wind speed behavior is faithfully reproduced. Good results were obtained. In fact, hourly wind speed data prove sufficient to reproduce the main statistical characteristics of wind speed: monthly mean, standard deviation, high hourly autocorrelation and persistence. This simple model is, therefore, easily adaptable to the study of any wind energy conversion system or to mixed power system planning and reliability studies.

Univariate Box-Jenkins time-series analysis has been used for modeling and forecasting monthly domestic electric energy consumption in the Eastern Province of Saudi Arabia. Autoregressive integrated moving average (ARIMA) models were developed using data for 5 yr and evaluated on forecasting new data for the sixth year. The optimum model derived is a multiplicative combination of seasonal and nonseasonal autoregressive parts, each being of the first order, following first differencing at both the seasonal and nonseasonal levels. Compared to regression and abductive network machine-learning models previously developed on the same data, ARIMA models require less data, have fewer coefficients, and are more accurate. The optimum ARIMA model forecasts monthly data for the evaluation year with an average percentage error of 3.8% compared to 8.1% and 5.6% for the best multiple-series regression and abductory induction mechanism (AIM) models, respectively; the mean-square forecasting error is reduced with the ARIMA model by factors of 3.2 and 1.6, respectively.

We address a problem faced by every supplier of electricity, i.e. forecasting the short-term electricity consumption. The introduction of new techniques has often been justified by invoking the nonlinearity of the problem. Our focus is directed to the question of deciding whether the problem is indeed nonlinear. First, we introduce a nonlinear measure of statistical dependence. Second, we analyse the linear and the nonlinear autocorrelation functions of the Czech electric consumption. Third, we compare the predictions of nonlinear models (artificial neural networks) with linear models (of the ARMA type). The correlational analysis suggests that forecasting the short-term evolution of the Czech electric load is primarily a linear problem. This is confirmed by the comparison of the predictions. In the light of this case study, the conditions under which neural networks could be superior to linear models are discussed.

In recent years the singular spectrum analysis (SSA) technique has been further developed and applied to many practical problems. The aim of this research is to extend and apply the SSA method, using the UK Industrial Production series. The performance of the SSA and multivariate SSA (MSSA) techniques was assessed by applying it to eight series measuring the monthly seasonally unadjusted industrial production for the main sectors of the UK economy. The results are compared with those obtained using the autoregressive integrated moving average and vector autoregressive models.
We also develop the concept of causal relationship between two time series based on the SSA techniques. We introduce several criteria which characterize this causality. The criteria and tests are based on the forecasting accuracy and predictability of the direction of change. The proposed tests are then applied and examined using the UK industrial production series. Copyright

We use univariate and multivariate singular spectrum analyses to predict the inflation rate as well as changes in the direction of inflation time series for the United States. We use consumer price indices and real-time chain-weighted GDP price index series in these prediction exercises. Moreover, we compare our out-of-sample, $h$-step-ahead moving prediction results with other prediction results based on methods such as activity-based NAIRU Phillips curve, $AR(p)$, the dynamic factors model, and random walk models with the latter as a naive forecasting method. We use short-run (quarterly) and long-run (one to six years) time windows for predictions and find that multivariate singular spectrum analysis outperforms all other competing prediction methods. Also, we confirm the results of earlier studies that prediction of the inflation rate in the United States during the period of the ``Great Moderation" is less challenging compared to the more volatile inflationary period of 1970-1985.

a b s t r a c t The present study applies three time series models, namely, Grey-Markov model, Grey-Model with rolling mechanism, and singular spectrum analysis (SSA) to forecast the consumption of conventional energy in India. Grey-Markov model has been employed to forecast crude-petroleum consumption while Grey-Model with rolling mechanism to forecast coal, electricity (in utilities) consumption and SSA to predict natural gas consumption. The models for each time series has been selected by carefully examining the structure of the individual time series. The mean absolute percentage errors (MAPE) for two out of sample forecasts have been obtained as follows: 1.6% for crude-petroleum, 3.5% for coal, 3.4% for electricity and 3.4% for natural gas consumption. For two out of sample forecasts, the prediction accuracy for coal consumption was 97.9%, 95.4% while for electricity consumption the prediction accu-racy was 96.9%, 95.1%. Similarly, the prediction accuracy for crude-petroleum consumption was found to be 99.2%, 97.6% while for natural gas consumption these values were 98.6%, 94.5%. The results obtained have also been compared with those of Planning Commission of India's projection. The comparison clearly points to the enormous potential that these time series models possess in energy consumption forecasting and can be considered as a viable alternative.

This paper considers univariate online electricity demand forecasting for lead times from a half-hour-ahead to a day-ahead. A time series of demand recorded at half-hourly intervals contains more than one seasonal pattern. A within-day seasonal cycle is apparent from the similarity of the demand profile from one day to the next, and a within-week seasonal cycle is evident when one compares the demand on the corresponding day of adjacent weeks. There is strong appeal in using a forecasting method that is able to capture both seasonalities. The multiplicative seasonal ARIMA model has been adapted for this purpose. In this paper, we adapt the Holt–Winters exponential smoothing formulation so that it can accommodate two seasonalities. We correct for residual autocorrelation using a simple autoregressive model. The forecasts produced by the new double seasonal Holt–Winters method outperform those from traditional Holt–Winters and from a well-specified multiplicative double seasonal ARIMA model.

Artificial neural network models were developed to forecast South Korea's transport energy demand. Various independent variables, such as GDP, population, oil price, number of vehicle registrations, and passenger transport amount, were considered and several good models (Model 1 with GDP, population, and passenger transport amount; Model 2 with GDP, number of vehicle registrations, and passenger transport amount; and Model 3 with oil price, number of vehicle registrations, and passenger transport amount) were selected by comparing with multiple linear regression models. Although certain regression models obtained better R-squared values than neural network models, this does not guarantee the fact that the former is better than the latter because root mean squared errors of the former were much inferior to those of the latter. Also, certain regression model had structural weakness based on P-value. Instead, neural network models produced more robust results. Forecasted results using the neural network models show that South Korea will consume around 37 MTOE of transport energy in 2025.

This study estimates electricity demand functions for Sri Lanka using six econometric techniques. It shows that the preferred specifications differ somewhat and there is a wide range in the long-run price and income elasticities with the estimated long-run income elasticity ranging from 1.0 to 2.0 and the long-run price elasticity from 0 to −0.06. There is also a wide range of estimates of the speed with which consumers would adjust to any disequilibrium, although the estimated impact income elasticities tended to be more in agreement ranging from 1.8 to 2.0. Furthermore, the estimated effect of the underlying energy demand trend varies between the different techniques; ranging from being positive to zero to predominantly negative. Despite these differences, the forecasts generated from the six models up until 2025 do not differ significantly. It is therefore encouraging that the Sri Lanka electricity authorities can have some faith in econometrically estimated models used for forecasting. Nonetheless, by the end of the forecast period in 2025 there is a variation of around 452 MW in the base forecast peak demand that, in relative terms for a small electricity generation system like Sri Lanka's, represents a considerable difference.

This study investigates the short-run dynamics and long-run equilibrium relationship between residential electricity demand and factors influencing demand – per capita income, price of electricity, price of kerosene oil and price of liquefied petroleum gas – using annual data for Sri Lanka for the period, 1960–2007. The study uses unit root, cointegration and error-correction models. The long-run demand elasticities of income, own price and price of kerosene oil (substitute) were estimated to be 0.78, − 0.62, and 0.14 respectively. The short-run elasticities for the same variables were estimated to be 0.32, − 0.16 and 0.10 respectively. Liquefied petroleum (LP) gas is a substitute for electricity only in the short-run with an elasticity of 0.09. The main findings of the paper support the following (1) increasing the price of electricity is not the most effective tool to reduce electricity consumption (2) existing subsidies on electricity consumption can be removed without reducing government revenue (3) the long-run income elasticity of demand shows that any future increase in household incomes is likely to significantly increase the demand for electricity and (4) any power generation plans which consider only current per capita consumption and population growth should be revised taking into account the potential future income increases in order to avoid power shortages in the country.

This paper considers forecasting techniques to predict the 24 market-clearing prices of a day-ahead electric energy market. The techniques considered include time series analysis, neural networks and wavelets. Within the time series procedures, the techniques considered comprise ARIMA, dynamic regression and transfer function. Extensive analysis is conducted using data from the PJM Interconnection. Relevant conclusions are drawn on the effectiveness and flexibility of any one of the considered techniques. Furthermore, they are exhaustively compared among themselves.

An artificial neural network model is developed to relate the electric energy consumption in the Eastern Province of Saudi Arabia to the weather data (temperature and humidity), global solar radiation and population. A two layered feedforward neural network is used for the modelling. The inputs to the neural network are the independent variables and the output is the electric energy consumption. Seven years' of data are used for model building and validation. Model adequacy is established by a visual inspection technique and the chi-square test. Model validation, which reflects the suitability of the model for future predictions is performed by comparing the predictions of the model with future data that was not used for model building. Comparison with a regression model shows that the neural network model performs better for predictions.

In the early 2000s, the Republic of Turkey has initiated an ambitious reform program in her electricity market, which requires privatization, liberalization as well as a radical restructuring. The most controversial reason behind, or justification for, recent reforms has been the rapid electricity demand growth; that is to say, the whole reform process has been a part of the endeavors to avoid the so-called “energy crisis”. Using cointegration analysis and autoregressive integrated moving average (ARIMA) modelling, the present article focuses on this issue by both providing an electricity demand estimation and forecast, and comparing the results with official projections. The study concludes, first, that consumers’ respond to price and income changes is quite limited and therefore there is a need for economic regulation in Turkish electricity market; and second, that the current official electricity demand projections highly overestimate the electricity demand, which may endanger the development of both a coherent energy policy in general and a healthy electricity market in particular.

The influence of economic and demographic variables on the annual electricity consumption in Italy has been investigated with the intention to develop a long-term consumption forecasting model.The time period considered for the historical data is from 1970 to 2007. Different regression models were developed, using historical electricity consumption, gross domestic product (GDP), gross domestic product per capita (GDP per capita) and population.A first part of the paper considers the estimation of GDP, price and GDP per capita elasticities of domestic and non-domestic electricity consumption. The domestic and non-domestic short run price elasticities are found to be both approximately equal to −0.06, while long run elasticities are equal to −0.24 and −0.09, respectively. On the contrary, the elasticities of GDP and GDP per capita present higher values.In the second part of the paper, different regression models, based on co-integrated or stationary data, are presented. Different statistical tests are employed to check the validity of the proposed models.A comparison with national forecasts, based on complex econometric models, such as Markal-Time, was performed, showing that the developed regressions are congruent with the official projections, with deviations of ±1% for the best case and ±11% for the worst. These deviations are to be considered acceptable in relation to the time span taken into account.

We address a problem faced by every supplier of electricity, i.e. forecasting the short-term electricity consumption. The introduction of new techniques has often been justified by invoking the nonlinearity of the problem. Our focus is directed to the question of deciding whether the problem is indeed nonlinear. First, we introduce a nonlinear measure of statistical dependence. Second, we analyse the linear and the nonlinear autocorrelation functions of the Czech electric consumption. Third, we compare the predictions of nonlinear models (artificial neural networks) with linear models (of the ARMA type). The correlational analysis suggests that forecasting the short-term evolution of the Czech electric load is primarily a linear problem. This is confirmed by the comparison of the predictions. In the light of this case study, the conditions under which neural networks could be superior to linear models are discussed.

Given two sources of forecasts of the same quantity, it is possible to compare prediction records. In particular, it can be useful to test the hypothesis of equal accuracy in forecast performance. We analyse the behaviour of two possible tests, and of modifications of these tests designed to circumvent shortcomings in the original formulations. As a result of this analysis, a recommendation for one particular testing approach is made for practical applications.

We propose a test of the null hypothesis that an observable series is stationary around a deterministic trend. The series is expressed as the sum of deterministic trend, random walk, and stationary error, and the test is the LM test of the hypothesis that the random walk has zero variance. The asymptotic distribution of the statistic is derived under the null and under the alternative that the series is difference-stationary. Finite sample size and power are considered in a Monte Carlo experiment. The test is applied to the Nelson-Plosser data, and for many of these series the hypothesis of trend stationarity cannot be rejected.

Because South Korea's industries depend heavily on imported energy sources (fifth largest importer of oil and second largest importer of liquefied natural gas in the world), the accurate estimating of its energy demand is critical in energy policy-making. This research proposes an artificial neural network model (a structure with feed-forward multilayer perceptron, error back-propagation algorithm, momentum process, and scaled data) to efficiently estimate the energy demand for South Korea. The model has four independent variables, such as gross domestic product (GDP), population, import, and export amounts. The data are obtained from diverse local and international sources. The proposed model better estimated energy demand than a linear regression model (a structure with multiple linear variables and least square method) or an exponential model (a structure with mixed integer variables, branch and bound method, and Broyden–Fletcher–Goldfarb–Shanno (BFGS) method) in terms of root mean squared error (RMSE). The model also forecasted better than the other two models in terms of RMSE without any over-fitting problem. Further testing with four scenarios based upon reliable source data showed unanticipated results. Instead of growing permanently, the energy demands peaked at certain points, and then decreased gradually. This trend is quite different from the results by regression or exponential model.

Energy dependency (ED) implies the extent to which an economy relies upon imports in order to meet its energy needs. The ED is calculated as net imports divided by the sum of gross inland energy consumption plus bunkers. This study aims at obtaining numerical equations to estimate of Turkey's energy dependency based on basic energy indicators and sectoral energy consumption by using artificial neural network (ANN) technique. It seeks to contribute to the strategies necessary to preserve the supply–demand balance of Turkey. For this purpose, two different models were used to train the ANN approach. In Model 1, main energy indicators such as total production of primary energy per capita, total gross electricity generation per capita and final energy consumption per capita were used in the input layer of the ANN while sectoral energy consumption per capita was used in Model 2.The ED was in the output layer for both models. Different models were employed to estimate the ED with a high confidence for future projections. The R2 values of ED were found to be 0.999 for both models. In accordance with the analysis results, ED is expected to increase from 72% to 82% within 14 years of period. Consequently, the utilization of renewable energy sources and nuclear energy is strictly recommended to ensure the ED stability in Turkey.

In this study, we used ARIMA, seasonal ARIMA (SARIMA) and alternatively the regression model with seasonal latent variable in forecasting electricity demand by using data that belongs to "Kayseri and Vicinity Electricity Joint-Stock Company" over the 1997:1-2005:12 periods. This study tries to examine the advantages of forecasting with ARIMA, SARIMA methods and with the model has seasonal latent variable to each other. The results support that ARIMA and SARIMA models are unsuccessful in forecasting electricity demand. The regression model with seasonal latent variable used in this study gives more successful results than ARIMA and SARIMA models because also this model can consider seasonal fluctuations and structural breaks.

Online short-term load forecasting is needed for the real-time scheduling of electricity generation. Univariate methods have been developed that model the intraweek and intraday seasonal cycles in intraday load data. Three such methods, shown to be competitive in recent empirical studies, are double seasonal ARMA, an adaptation of Holt-Winters exponential smoothing for double seasonality, and another, recently proposed, exponential smoothing method. In multiple years of load data, in addition to intraday and intraweek cycles, an intrayear seasonal cycle is also apparent. We extend the three double seasonal methods in order to accommodate the intrayear seasonal cycle. Using six years of British and French data, we show that for prediction up to a day-ahead the triple seasonal methods outperform the double seasonal methods, and also a univariate neural network approach. Further improvement in accuracy is produced by using a combination of the forecasts from two of the triple seasonal methods.

In 2002 the Government of Sri Lanka proposed power sector policy guidelines for the first time in its history in order to facilitate the restructuring of the sector. This paper attempts to critically examine and appraise the Government’s proposals with suggestions for improvements. The methodology employed is to first examine the requirements of the Sri Lankan power sector by analysing the current problems that the power sector faces and to empirically estimate electricity demand to identify the future consumption and capacity expansion needs of the sector. Secondly, it is assessed to what extent the proposed reforms address the requirements of the sector identified above. Finally, alternative proposals are introduced in order to address the identified flaws in the current proposed reforms.

This paper, has tried to forecast monthly maximum electricity demand for the state Maharashtra, India, using Multiplicative Seasonal Autoregressive Integrated Moving Average (MSARIMA) method for seasonally unadjusted monthly data spanning from April 1980 to June 1999. The forecasted period is 18 months ahead from June 1999. This study's basic findings are that the series does not reveal any drastic change for the forecasted period. It continues to follow the same trend along with the seasonal variation. Copyright 2002 by Taylor and Francis Group