Dataset Information

Empirical prediction intervals improve energy forecasting.

ABSTRACT: Hundreds of organizations and analysts use energy projections, such as those contained in the US Energy Information Administration (EIA)'s Annual Energy Outlook (AEO), for investment and policy decisions. Retrospective analyses of past AEO projections have shown that observed values can differ from the projection by several hundred percent, and thus a thorough treatment of uncertainty is essential. We evaluate the out-of-sample forecasting performance of several empirical density forecasting methods, using the continuous ranked probability score (CRPS). The analysis confirms that a Gaussian density, estimated on past forecasting errors, gives comparatively accurate uncertainty estimates over a variety of energy quantities in the AEO, in particular outperforming scenario projections provided in the AEO. We report probabilistic uncertainties for 18 core quantities of the AEO 2016 projections. Our work frames how to produce, evaluate, and rank probabilistic forecasts in this setting. We propose a log transformation of forecast errors for price projections and a modified nonparametric empirical density forecasting method. Our findings give guidance on how to evaluate and communicate uncertainty in future energy outlooks.

SUBMITTER: Kaack LH

PROVIDER: S-EPMC5565406 | biostudies-literature | 2017 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Empirical prediction intervals improve energy forecasting.

Kaack Lynn H LH Apt Jay J Morgan M Granger MG McSharry Patrick P

Proceedings of the National Academy of Sciences of the United States of America 20170731 33

Hundreds of organizations and analysts use energy projections, such as those contained in the US Energy Information Administration (EIA)'s Annual Energy Outlook (AEO), for investment and policy decisions. Retrospective analyses of past AEO projections have shown that observed values can differ from the projection by several hundred percent, and thus a thorough treatment of uncertainty is essential. We evaluate the out-of-sample forecasting performance of several empirical density forecasting met ...[more]

PMID: 28760997

Similar Datasets

Project description:IntroductionSuicide is a major social and health issue in India. Yearly statistics show a concerning increasing pattern of suicidal deaths in India which is higher in comparison to the global trend. There is limited evidence regarding historical analysis of suicide or any forecasting for suicide in India towards predicting the possible risks of death due to suicide.MethodsThis paper examines the trend of suicide rate and characteristics of suicide victims in India, based on the longitudinal time series data over the last 50 years-collected from the National Crime Record Bureau Reports (1969 to 2018) of the Government of India. In our analysis, we have used the time series model to forecast the suicide rates in India for the next decade. ARIMA (4,1,0) model is found to be the best fit model for forecasting the data.FindingsThere has been an observable and rising trend of suicide rates in India over the last five decades. The forecast indicates a continuance of rising suicide cases for an upcoming couple of years in India with a limited decline in the following years. The prediction model indicates a future relatively consistent pattern of suicide in India which does not seem to be a very encouraging trend. As we have not included the period staring the year 2020 onwards affected by Covid-19 and which has several disruptions in personal and family spaces, the projected suicide trend during the period of next two to three years (2020-22) may rise far high and then it may show a declining path. Along with this, there is a shift in means of suicide in the last couple of decades. Constituting the second-highest number of cases, Illness associated suicide was visibly a serious concern.ConclusionThe present analysis finds that there is no visible substantial relief for suicide deaths during the coming years in India. On the other hand, more extensive exploration of sample cases may provide important information for suicide prevention. Availability of detailed and more inclusive data will be highly useful for analysis and suicide preventive policies. Investment in public health care and other welfare activities like education and employment generation will yield visible positive results in suicide control.

Dataset Information

Empirical prediction intervals improve energy forecasting.

Publications

Empirical prediction intervals improve energy forecasting.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets