Previous Forecasts Research Articles

Determination of Optimum Smoothing Constant of Single Exponential Smoothing Method: A Case Study

Exponential smoothing is a sophisticated forecasting method that works based on previous forecast plus a percentage of the forecast error. A key issue of this technique is the proper choice of exponential smoothing constant. In order to minimize forecasting errors, choosing an appropriate value of smoothing constant is very crucial. In this study, a framework is developed for the selection of optimal value of smoothing constant that minimizes a measure of forecast errors like mean square error (MSE) and mean absolute deviation (MAD). Both “trial & error” and Excel based non-linear optimizer (“Excel Solver”) are used for this purpose. To validate the proposed model, necessary demand data of Ruchi Jhal Muri from years 2010-2016 from Square Food & Beverage Ltd. in Mohakhali, Dhaka were collected. The optimum values of smoothing constant under trial & error method are 0.31 and 0.14 for minimum MAD and MSE respectively whereas for excel solver, values are 0.314 and 0.143 with respect to minimum MAD and MSE. Although both methods provide approximately the same results but excel solver is much easier & requires less time for deriving optimum solution. This study will provide an outline for the forecast planners as well as manufacturers to improve the accuracy of exponential forecasting through using Excel Solver for determining the optimum value of smoothing constant.

The Representation of Tropospheric Water Vapor Over Low-Latitude Oceans in (Re-)analysis: Errors, Impacts, and the Ability to Exploit Current and Prospective Observations

This paper addresses the representation of lower tropospheric water vapor in the meteorological analyses—fully detailed estimates of atmospheric state—providing the wide temporal and spatial coverage used in many process studies. Analyses are produced in a cycle combining short forecasts from initial conditions with data assimilation that optimally estimates the state of the atmosphere from the previous forecasts and new observations, providing initial conditions for the next set of forecasts. Estimates of water vapor are among the less certain aspects of the state because the quantity poses special challenges for data assimilation while being particularly sensitive to the details of model parameterizations. Over remote tropical oceans observations of water vapor come from two sources: passive observations at microwave or infrared wavelengths that provide relatively strong constraints over large areas on column-integrated moisture but relatively coarse vertical resolution, and occultations of Global Positioning System provide much higher accuracy and vertical resolution but are relatively spatially coarse. Over low-latitude oceans, experiences with two systems suggest that current analyses reproduce much of the large-scale variability in integrated water vapor but have systematic errors in the representation of the boundary layer with compensating errors in the free troposphere; these errors introduce errors of order 10% in radiative heating rates through the free troposphere. New observations, such as might be obtained by future observing systems, improve the estimates of water vapor but this improvement is lost relatively quickly, suggesting that exploiting better observations will require targeted improvements to global forecast models.

The SDSS-IV extended Baryon Oscillation Spectroscopic Survey: final emission line galaxy target selection

We describe the algorithm used to select the Emission Line Galaxy (ELG) sample at $z \sim 0.85$ for the extended Baryon Oscillation Spectroscopic Survey of the Sloan Digital Sky Survey IV, using photometric data from the DECam Legacy Survey. Our selection is based on a selection box in the $g-r$ vs. $r-z$ colour-colour space and a cut on the $g$-band magnitude, to favour galaxies in the desired redshift range with strong [OII] emission. It provides a target density of 200 deg$^{-2}$ on the North Galactic Cap (NGC) and of 240 deg$^{-2}$ on the South Galactic Cap (SGC), where we use a larger selection box because of deeper imaging. We demonstrate that this selection passes the eBOSS requirements in terms of homogeneity. About 50,000 ELGs have been observed since the observations have started in 2016, September. These roughly match the expected redshift distribution, though the measured efficiency is slightly lower than expected. The efficiency can be increased by enlarging the redshift range and with incoming pipeline improvement. The cosmological forecast based on these first data predict $\sigma_{D_V}/D_V = 0.023$, in agreement with previous forecasts. Lastly, we present the stellar population properties of the ELG SGC sample. Once observations are completed, this sample will be suited to provide a cosmological analysis at $z \sim 0.85$, and will pave the way for the next decade of massive spectroscopic cosmological surveys, which heavily rely on ELGs. The target catalogue over the SGC will be released along with DR14.

2014 Election forecast - a post-election analysis

General elections are held every five years in South Africa. During the 12 to 24 hour period after the close of the voting booths, the expected final results are of huge interest to the electorate and politicians. In the past, the Council for Scientific and Industrial Research (CSIR) has developed an election forecasting model in order to provide the media and political analysts with forecasts of the final results during this period of peak interest. In formulating this model, which forecasts the election results as the results from voting districts (VDs) become available, some assumptions had to be made. In particular, assumptions were made about the clustering of previous voting patterns as well as the order in which VD results are released. This election forecasting model had been used successfully for a number of elections in the past and in these previous elections, with around 5%-10% of the results available, the predictions produced by the model were very close to the final outcome, particularly for the ANC, being the largest party. For the 2014 national election, however, the predictions, with close to 50% of the voting district results known (equivalent to an estimated 40% of the total votes), were still not accurate and varied by more than 1% for both the ANC and the EFF. This paper outlines a post-election analysis to determine the reasons for these discrepancies and how they relate directly to the model assumptions. The aim is to highlight how practical realities can affect the assumptions and consequently their impact on the forecasted results. Reference is made to previous election forecasts and the 2014 post-election analysis is presented.

Online Multi-Task Learning Framework for Ensemble Forecasting

Ensemble forecasting is a widely-used numerical prediction method for modeling the evolution of nonlinear dynamic systems. To predict the future state of such systems, a set of ensemble member forecasts is generated from multiple runs of computer models, where each run is obtained by perturbing the starting condition or using a different model representation of the system. The ensemble mean or median is typically chosen as a point estimate for the ensemble member forecasts. These approaches are limited in that they assume each ensemble member is equally skillful and may not preserve the temporal autocorrelation of the predicted time series. To overcome these limitations, we present an online multi-task learning framework called ORION to estimate the optimal weights for combining the ensemble member forecasts. Unlike other existing formulations, the proposed framework is novel in that its learning algorithm must backtrack and revise its previous forecasts before making future predictions if the earlier forecasts were incorrect when verified against new observation data. We termed this strategy as online learning with restart. Our proposed framework employs a graph Laplacian regularizer to ensure consistency of the predicted time series. It can also accommodate different types of loss functions, including ϵ-insensitive and quantile loss functions, the latter of which is particularly useful for extreme value prediction. A theoretical proof demonstrating the convergence of our algorithm is also given. Experimental results on seasonal soil moisture forecasts from 12 major river basins in North America demonstrate the superiority of ORION compared to other baseline algorithms.

Forecasted trends in disability and life expectancy in England and Wales up to 2025: a modelling study

SummaryBackgroundReliable estimation of future trends in life expectancy and the burden of disability is crucial for ageing societies. Previous forecasts have not considered the potential impact of trends in disease incidence. The present prediction model combines population trends in cardiovascular disease, dementia, disability, and mortality to forecast trends in life expectancy and the burden of disability in England and Wales up to 2025.MethodsWe developed and validated the IMPACT-Better Ageing Model—a probabilistic model that tracks the population aged 35–100 years through ten health states characterised by the presence or absence of cardiovascular disease, dementia, disability (difficulty with one or more activities of daily living) or death up to 2025, by use of evidence-based age-specific, sex-specific, and year-specific transition probabilities. As shown in the English Longitudinal Study of Ageing, we projected continuing declines in dementia incidence (2·7% per annum), cardiovascular incidence, and mortality. The model estimates disability prevalence and disabled and disability-free life expectancy by year.FindingsBetween 2015 and 2025, the number of people aged 65 years and older will increase by 19·4% (95% uncertainty interval [UI] 17·7–20·9), from 10·4 million (10·37–10·41 million) to 12·4 million (12·23–12·57 million). The number living with disability will increase by 25·0% (95% UI 21·3–28·2), from 2·25 million (2·24–2·27 million) to 2·81 million (2·72–2·89 million). The age-standardised prevalence of disability among this population will remain constant, at 21·7% (95% UI 21·5–21·8) in 2015 and 21·6% (21·3–21·8) in 2025. Total life expectancy at age 65 years will increase by 1·7 years (95% UI 0·1–3·6), from 20·1 years (19·9–20·3) to 21·8 years (20·2–23·6). Disability-free life expectancy at age 65 years will increase by 1·0 years (95% UI 0·1–1·9), from 15·4 years (15·3–15·5) to 16·4 years (15·5–17·3). However, life expectancy with disability will increase more in relative terms, with an increase of roughly 15% from 2015 (4·7 years, 95% UI 4·6–4·8) to 2025 (5·4 years, 4·7–6·4).InterpretationThe number of older people with care needs will expand by 25% by 2025, mainly reflecting population ageing rather than an increase in prevalence of disability. Lifespans will increase further in the next decade, but a quarter of life expectancy at age 65 years will involve disability.FundingBritish Heart Foundation.

Ensemble flood forecasting based on operational forecasts of the regional Eta EPS in the Taquari-Antas basin

ABSTRACT Hydrological Ensemble Prediction Systems (HEPS) play an important role on operational flood forecasting. Unlike in deterministic approach, which relies on a single prediction of future river flows, these systems can represent the forecast uncertainty and provide a better detection of extreme hydro-meteorological events. In this context, the present study aimed to assess both the quality of ensemble flood forecasts on Taquari-Antas basin and its potential to provide additional information to a local Flood Alert System. The hydrological model MGB-IPH was coupled to the high-resolution meteorological EPS Eta model with five members of different parameterization schemes and boundary conditions, as well as to the deterministic version of Eta regional model. On a single event evaluation, the peak discharge was reasonable well predicted by at least one ensemble member, in nearly all forecasts, with a good prediction of the flood timing for the considered lead times. In a comparison with deterministic forecasts, the ensemble ones showed higher accuracy and higher probability of detection (POD) for the reference thresholds, preserving false alarm rates at reasonably low levels. An overall tendency of underestimation was also identified, with most observations falling between the higher ranks of the ensemble. Furthermore, the combination of previous forecasts (t-12h) with the recent ones leads to a slight increase of ensemble spread and POD, despite the performance reduction in terms of accuracy and bias for the ensemble mean. Results suggest that there is a benefit in having hydrological ensemble forecasts obtained from the high resolution EPS Eta model, which can be used as a complementary information to a local Flood Alert System supporting pre-alert issues and Civil Defense internal planning actions.

Time-delay cosmography: increased leverage with angular diameter distances

Strong lensing time-delay systems constrain cosmological parameters via the so-called time-delay distance and the angular diameter distance to the lens. In previous studies, only the former information was used in forecasting cosmographic constraints. In this paper, we show that the cosmological constraints improve significantly when the latter information is also included. Specifically, the angular diameter distance plays a crucial role in breaking the degeneracy between the curvature of the Universe and the time-varying equation of state of dark energy. Using a mock sample of 55 bright quadruple lens systems based on expectations for ongoing/future imaging surveys, we find that adding the angular diameter distance information to the time-delay distance information and the Planck's measurements of the cosmic microwave background anisotropies improves the constraint on the constant equation of state by 30%, on the time variation in the equation of state by a factor of two, and on the Hubble constant in the flat ΛCDM model by a factor of two. Therefore, previous forecasts for the statistical power of time-delay systems were overly pessimistic, i.e., time-delay systems are more powerful than previously appreciated.

Constraining cosmology and ionization history with combined 21 cm power spectrum and global signal measurements

Improvements in current instruments and the advent of next-generation instruments will soon push observational 21 cm cosmology into a new era, with high significance measurements of both the power spectrum and the mean ("global") signal of the 21 cm brightness temperature. In this paper we use the recently commenced Hydrogen Epoch of Reionization Array as a worked example to provide forecasts on astrophysical and cosmological parameter constraints. In doing so we improve upon previous forecasts in a number of ways. First, we provide updated forecasts using the latest best-fit cosmological parameters from the Planck satellite, exploring the impact of different Planck datasets on 21 cm experiments. We also show that despite the exquisite constraints that other probes have placed on cosmological parameters, the remaining uncertainties are still large enough to have a non-negligible impact on upcoming 21 cm data analyses. While this complicates high-precision constraints on reionization models, it provides an avenue for 21 cm reionization measurements to constrain cosmology. We additionally forecast HERA's ability to measure the ionization history using a combination of power spectrum measurements and semi-analytic simulations. Finally, we consider ways in which 21 cm global signal and power spectrum measurements can be combined, and propose a method by which power spectrum results can be used to train a compact parameterization of the global signal. This parameterization reduces the number of parameters needed to describe the global signal, increasing the likelihood of a high significance measurement.

USD SGD Outlook-2016

Interest rate:1. Historically, Singapore has had a lower interest rate compared to US, indicating potential appreciation in SGD against USD. But that leading edge over USD is gone now for SGD since last 3-4 years.GDP:1. Singapore economy is bracing itself for sub-3% annual growth rate in coming years. It would be first time in last 6 years that economy will post 2% growth rate in 2015 (it has slashed its previous forecast of 2 to 4% annual growth) .2. Economy is slowing due to various external reasons such as China’s deceleration, weaker trade and poor growth in ASEAN region, ongoing and impending US rate hikes, Oil price slump, etc.3. We need to be reminded of the fact that when China's growth plunged from 9.5% in 2011 to 7.7% in 2012, the rate of growth here also dropped by 50% over the same period, from 6.2% to 3.4%. Though, such a decline is not expected in future but the country is vulnerable to China’s economic slowdown.Inflation:1. MAS now expects inflation at 0.5% this year (2015) compared to its previous forecast of 0.5-1.5%.2. MAS has admitted that wage inflation is still high.3. In November 2015, Country has witnessed 13th straight months of negative inflation rate (the longest streak since 1987 crisis). It is giving an indication that monetary easing in the offing in Q1 2016.Singapore dollar nominal effective exchange rate (NEER):1. Citing continued SGD depreciation since last 3 years and potential depreciation in the future and growth slowdown, there is likelihood that the MAS will review its current view towards NEER and ease the policy by changing the currency’s appreciatory slope to neutral, from its current slightly appreciatory bias. Though, any aggressive monetary policy loosening is not expected citing the impending US Federal Reserve Rate hikes in 2016.2. SGD exchange rate has moved towards the bottom of MAS policy band, suggesting even more loosening of monetary policy for the economy is on the cards.

Farmers’ Perception and Willingness to Pay for Climate Information in Burkina Faso

<p>This study highlights the perception and the option value of climate information in the sahelian and sudano-sahelian agro-climatic zones of Burkina Faso. From the sampled farmers only 21.78% have access to climate information, while 93% of them need climate information to guide through their decision in planning agricultural activities. Option value shows the interest granted by farmers using climate information. About 64% of the farmers are willing to pay an average of Francs CFA 546.34 to obtain climate information. The willingness to pay is determined by the the farmers’ ability to predict the climate, to use radio as a means of information, farmers’ awareness of previous forecasts and early onset of rainy season. While farmers considered benefits from the use of climate information, it is clear that its contribution to farm income remains a field of research to explore. Thus it is necessary to experiment with individual farms and to evaluate the contribution of climate information to the value added in different crops and farmers’ income.<strong></strong></p>

How accurate are forecasts of costs of energy? A methodological contribution

Forecasts of the cost of energy are typically presented as point estimates; however forecasts are seldom accurate, which makes it important to understand the uncertainty around these point estimates. The scale of the differences between forecasts and outturns (i.e. contemporary estimates) of costs may have important implications for government decisions on the appropriate form (and level) of support, modelling energy scenarios or industry investment appraisal. This paper proposes a methodology to assess the accuracy of cost forecasts. We apply this to levelised costs of energy for different generation technologies due to the availability of comparable forecasts and contemporary estimates, however the same methodology could be applied to the components of levelised costs, such as capital costs. The estimated “forecast errors” capture the accuracy of previous forecasts and can provide objective bounds to the range around current forecasts for such costs. The results from applying this method are illustrated using publicly available data for on- and off-shore wind, Nuclear and CCGT technologies, revealing the possible scale of “forecast errors” for these technologies.

Accounting for groundwater in stream fish thermal habitat responses to climate change.

Forecasting climate change effects on aquatic fauna and their habitat requires an understanding of how water temperature responds to changing air temperature (i.e., thermal sensitivity). Previous efforts to forecast climate effects on brook trout (Salvelinus fontinalis) habitat have generally assumed uniform air-water temperature relationships over large areas that cannot account for groundwater inputs and other processes that operate at finer spatial scales. We developed regression models that accounted for groundwater influences on thermal sensitivity from measured air-water temperature relationships within forested watersheds in eastern North America (Shenandoah National Park, Virginia, USA, 78 sites in nine watersheds). We used these reach-scale models to forecast climate change effects on stream temperature and brook trout thermal habitat, and compared our results to previous forecasts based upon large-scale models. Observed stream temperatures were generally less sensitive to air temperature than previously assumed, and we attribute this to the moderating effect of shallow groundwater inputs. Predicted groundwater temperatures from air-water regression models corresponded well to observed groundwater temperatures elsewhere in the study area. Predictions of brook trout future habitat loss derived from our fine-grained models. were far less pessimistic than those from prior models developed at coarser spatial resolutions. However, our models also revealed spatial variation in thermal sensitivity within and among catchments resulting in a patchy distribution of thermally suitable habitat. Habitat fragmentation due to thermal barriers therefore may have an increasingly important role for trout population viability in headwater streams. Our results demonstrate that simple adjustments to air-water temperature regression models can provide a powerful and cost-effective approach for predicting future stream temperatures while accounting for effects of groundwater.

Development of the Model of Dynamic Storage Distribution in Data Processing Centers

The paper reviews dynamic distribution of storage resources among the users in data processing centers. The process of changing memory usage state was revealed to be the process of Markov. The paper proposes the development of stochastic model of the memory and computing usage distribution and the development of probability density functions over practical data. Parameters of probability density functions were defined with the help of stochastic model and practical data. The calculation of the developed model and the parameters of the probability density function is realized dynamically during the ongoing process. At the beginning of each time interval, it is forecasted that the process will be shifted to which state with which maximum probability. The adequacy of the previous forecasts is monitored. Note that, over the time, the quality of the forecast and the level of adequacy increases. The model is used in the virtualization of storage resources usage process and ensures the use of storage resources without wasting. Structure of visualization base is given. The base enables to monitor all stages of the process. Using monitoring base the issues can be resolved to analyze different aspects of the process. Recommendations are given on the use of obtained results.

The Regional Ice Prediction System (RIPS): verification of forecast sea ice concentration

In recent years, the demand for improved environmental forecasts in the Arctic has intensified as maritime transport and offshore exploration increase. As a result, Canada has accepted responsibility for the preparation and issuing services for the new Arctic MET/NAV Areas XVII and XVIII. Environmental forecasts are being developed based on a new integrated Arctic marine prediction system. Here, we present the first phase of this initiative, a short‐term pan‐Arctic 1/12° resolution Regional Ice Prediction System (RIPS). RIPS is currently set to perform four 48 h forecasts per day. The RIPS forecast model (CICE 4.0) is forced by atmospheric forecasts from the Environment Canada regional deterministic prediction system. It is initialized with a 3D‐Var analysis of sea ice concentration and the ice velocity field and thickness distribution from the previous forecast. The other forcing (surface current) and initialization fields (mixed‐layer depth, sea surface temperature and salinity) come from the 1/4° resolution Global Ice Ocean Prediction System. Three verification methods for sea ice concentration are presented. Overall, verifications over a complete seasonal cycle (2011) against the Ice Mapping System ice extent product show that RIPS 48 h forecasts are better than persistence during the growth season while they have a lower skill than persistence during the melt period. A better representation of landfast ice, oceanic processes (wave–ice interactions, upwelling events, etc.) in the marginal ice zone and better initializing fields should lead to improved forecasts.

The validity of long-term prediction of quasi-biennial oscillation (QBO) as a proof of the exact seasonal synchronization of the equatorial stratospheric QBO cycle

Sixty years of radiosonde data of the equatorial stratospheric zonal wind have been used to develop the method of long-term wind variations forecast based on the seasonal regularities of quasi-biennial oscillation (QBO) evolution. The wind changes during all easterly wind regime descents in 1953–2014 in the middle stratosphere (from 10 to 70hPa) have been analyzed elaborately by the experimental data. Arguments are presented which show that the delay of easterly downward propagation is observed during each QBO cycle, not just in some cycles, as is commonly believed. Moreover it is shown that the start and the end of all easterly delays (the stages of stagnation) are strongly linked to seasons of a year. It causes the discretely varying length of the stagnation stages and the clearly defined quantized QBO cycle period which can only be equal to 24, 30 or 36 months. These regularities allow developing the method of long-term QBO-cycle prediction. The verification of previous forecasts by comparing with actually observed zonal wind changes in 2005–2014 is presented. Also the QBO development for the following period as far as December 2018 is predicted. We show that the predictability of the wind QBO in the equatorial stratosphere for a long time interval (up to five years) is the result of the existence of a strong seasonal dependence of the QBO evolution. On the other hand, the correctness of these QBO predictions is a strong proof of the exact seasonal synchronization of the QBO cycle in the equatorial stratosphere.

Continuous short-term irradiance forecasts using sky images

We present a system for forecasting occlusions of the sun and the expected Global Horizontal Irradiance (GHI) for solar power plants. Our system uses non-rigid registration for detecting cloud motion and a Kalman filter to establish continuous forecasts for up to 10min. The optimal parameters of the system are determined through the use of the binary classification metrics Precision, Recall and F2 Score while evaluating the forecasting of occlusions. The Kalman filter and the use of a dense motion field instead of a global cloud speed prove to be key elements of the forecasting pipeline: by incorporating information from previous forecasts into the current one, a Kalman filtering facilitates forecasting times below 3min and the dense motion field enhances the accuracy of our forecasts. Our evaluation of the proposed approach on 15days of real world data collected in Kitzingen, Bavaria, Germany, produced a mean RMSE for forecasting GHI of (164±9)Wm−2.

The Longevity Prospects of Australian Seniors: An Evaluation of Forecast Method and Outcome

AbstractContinuing rapid changes in the level and pattern of mortality require that forecasts are available that are timely, relevant and reliable. This paper evaluates a previous forecast of the mortality and longevity of Australian seniors, both in terms of the validity of the chosen method – the Booth–Maindonald–Smith (BMS) variant of Lee–Carter – and the accuracy and reliability of the forecast itself. The validity of the method is assessed by a comprehensive review and evaluation of available methods, confirming BMS as the method of choice. The accuracy and reliability of the forecast is assessed by comparing it with actual experience and with a new forecast of period and cohort survival probabilities and life expectancies. The evaluation and the current forecast itself will inform the actuarial profession and wider industry in the areas of mortality and longevity risk as well as public debate and policy in population health and ageing.

Greenhouse Gases Emission in Latvia: Modelling and Forecasting

This paper presents a forecast analysis of production in different sectors of Latvia's manufacturing. It considers the results of investigation ordered by Ministry of Environment Protection. The forecasts of production growth in different sectors of manufacturig are calculated up to 2020. The comparison with previous forecasts is made.

Evaluation and update of Norwegian and Danish oil production forecasts and implications for Swedish oil import

This paper presents an updated historic oil production analysis as well as an updated future oil production forecast for Norway and Denmark. Previous forecasts conducted by academic and official agencies using a variety of methodologies are contrasted and their accuracy examined. The bottom-up field-by-field methodology is found to be precise in the short-term, as it deviates by less than 1% from actual production.The impact of declining oil production in the North Sea on Sweden is explored as a case study. The historic and future trends regarding Swedish oil imports are presented and their vulnerability assessed using the Herfindahl–Hirschman index.