Improving load forecasting accuracy through combination of best forecasts

Neural network (NN) models have been widely used in the literature for short-term load forecasting. Their popularity is mainly due to their excellent learning and approximation capability. However, their forecasting performance significantly depends on several factors including initializing parameters, training algorithm, and NN structure. To minimize negative effects of these factors, this paper proposes a practically simple, yet effective and an efficient method to combine forecasts generated by NN models. The proposed method includes three main phases: (i) training NNs with different structures, (ii) selecting best NN models based on their forecasting performance for a validation set, and (iii) combination of forecasts for selected best NNs. Forecast combination is performed through calculating the mean of forecasts generated by best NN models. The performance of the proposed method is examined using real world data set. Comparative studies demonstrate that the accuracy of combined forecasts is significantly superior to those obtained from individual NN models.

Aggregation of Pi-based forecast to enhance prediction accuracy

In contrast to point forecast, prediction interval-based neural network offers itself as an effective tool to quantify the uncertainty and disturbances that associated with process data. However, single best neural network (NN) does not always guarantee to predict better quality of forecast for different data sets or a whole range of data set. Literature reported that ensemble of NNs using forecast combination produces stable and consistence forecast than single best NN. In this work, a NNs ensemble procedure is introduced to construct better quality of Pis. Weighted averaging forecasts combination mechanism is employed to combine the Pi-based forecast. As the key contribution of this paper, a new Pi-based cost function is proposed to optimize the individual weights for NN in combination process. An optimization algorithm, named simulated annealing (SA) is used to minimize the PI-based cost function. Finally, the proposed method is examined in two different case studies and compared the results with the individual best NNs and available simple averaging Pis aggregating method. Simulation results demonstrated that the proposed method improved the quality of Pis than individual best NNs and simple averaging ensemble method.

Application of air quality combination forecasting to Bogota

The bulk of existing work on the statistical forecasting of air quality is based on either neural networks or linear regressions, which are both subject to important drawbacks. In particular, while neural networks are complicated and prone to in-sample overfitting, linear regressions are highly dependent on the specification of the regression function. The present paper shows how combining linear regression forecasts can be used to circumvent all of these problems. The usefulness of the proposed combination approach is verified using both Monte Carlo simulation and an extensive application to air quality in Bogota, one of the largest and most polluted cities in Latin America. © 2014 Elsevier Ltd.

A panel data approach to economic forecasting: the bias-corrected average forecast

In this paper, we propose a novel approach to econometric forecasting of stationary and ergodic time series within a panel-data framework. Our key element is to employ the (feasible) bias-corrected average forecast. Using panel-data sequential asymptotics we show that it is potentially superior to other techniques in several contexts. In particular, it is asymptotically equivalent to the conditional expectation, i.e., has an optimal limiting mean-squared error. We also develop a zeromean test for the average bias and discuss the forecast-combination puzzle in small and large samples. Monte-Carlo simulations are conducted to evaluate the performance of the feasible bias-corrected average forecast in finite samples. An empirical exercise based upon data from a well known survey is also presented. Overall, theoretical and empirical results show promise for the feasible bias-corrected average forecast.

A panel data approach to economic forecasting: the bias-corrected average forecast

In this paper, we propose a novel approach to econometric forecasting of stationary and ergodic time series within a panel-data framework. Our key element is to employ the bias-corrected average forecast. Using panel-data sequential asymptotics we show that it is potentially superior to other techniques in several contexts. In particular it delivers a zero-limiting mean-squared error if the number of forecasts and the number of post-sample time periods is sufficiently large. We also develop a zero-mean test for the average bias. Monte-Carlo simulations are conducted to evaluate the performance of this new technique in finite samples. An empirical exercise, based upon data from well known surveys is also presented. Overall, these results show promise for the bias-corrected average forecast.

A panel data approach to economic forecasting: the bias-corrected average forecast

In this paper, we propose a novel approach to econometric forecasting of stationary and ergodic time series within a panel-data framework. Our key element is to employ the (feasible) bias-corrected average forecast. Using panel-data sequential asymptotics we show that it is potentially superior to other techniques in several contexts. In particular, it is asymptotically equivalent to the conditional expectation, i.e., has an optimal limiting mean-squared error. We also develop a zeromean test for the average bias and discuss the forecast-combination puzzle in small and large samples. Monte-Carlo simulations are conducted to evaluate the performance of the feasible bias-corrected average forecast in finite samples. An empirical exercise, based upon data from a well known survey is also presented. Overall, these results show promise for the feasible bias-corrected average forecast.

Using common features to understand the behavior of metal-commodity prices and forecast them at different horizons

The objective of this article is to study (understand and forecast) spot metal price levels and changes at monthly, quarterly, and annual horizons. The data to be used consists of metal-commodity prices in a monthly frequency from 1957 to 2012 from the International Financial Statistics of the IMF on individual metal series. We will also employ the (relatively large) list of co-variates used in Welch and Goyal (2008) and in Hong and Yogo (2009) , which are available for download. Regarding short- and long-run comovement, we will apply the techniques and the tests proposed in the common-feature literature to build parsimonious VARs, which possibly entail quasi-structural relationships between different commodity prices and/or between a given commodity price and its potential demand determinants. These parsimonious VARs will be later used as forecasting models to be combined to yield metal-commodity prices optimal forecasts. Regarding out-of-sample forecasts, we will use a variety of models (linear and non-linear, single equation and multivariate) and a variety of co-variates to forecast the returns and prices of metal commodities. With the forecasts of a large number of models (N large) and a large number of time periods (T large), we will apply the techniques put forth by the common-feature literature on forecast combinations. The main contribution of this paper is to understand the short-run dynamics of metal prices. We show theoretically that there must be a positive correlation between metal-price variation and industrial-production variation if metal supply is held fixed in the short run when demand is optimally chosen taking into account optimal production for the industrial sector. This is simply a consequence of the derived-demand model for cost-minimizing firms. Our empirical evidence fully supports this theoretical result, with overwhelming evidence that cycles in metal prices are synchronized with those in industrial production. This evidence is stronger regarding the global economy but holds as well for the U.S. economy to a lesser degree. Regarding forecasting, we show that models incorporating (short-run) commoncycle restrictions perform better than unrestricted models, with an important role for industrial production as a predictor for metal-price variation. Still, in most cases, forecast combination techniques outperform individual models.

Using common features to understand the behavior of metal-commodity prices and forecast them at different horizons

The objective of this article is to study (understand and forecast) spot metal price levels and changes at monthly, quarterly, and annual frequencies. Data consists of metal-commodity prices at a monthly and quarterly frequencies from 1957 to 2012, extracted from the IFS, and annual data, provided from 1900-2010 by the U.S. Geological Survey (USGS). We also employ the (relatively large) list of co-variates used in Welch and Goyal (2008) and in Hong and Yogo (2009). We investigate short- and long-run comovement by applying the techniques and the tests proposed in the common-feature literature. One of the main contributions of this paper is to understand the short-run dynamics of metal prices. We show theoretically that there must be a positive correlation between metal-price variation and industrial-production variation if metal supply is held fixed in the short run when demand is optimally chosen taking into account optimal production for the industrial sector. This is simply a consequence of the derived-demand model for cost-minimizing firms. Our empirical evidence fully supports this theoretical result, with overwhelming evidence that cycles in metal prices are synchronized with those in industrial production. This evidence is stronger regarding the global economy but holds as well for the U.S. economy to a lesser degree. Regarding out-of-sample forecasts, our main contribution is to show the benefits of forecast-combination techniques, which outperform individual-model forecasts - including the random-walk model. We use a variety of models (linear and non-linear, single equation and multivariate) and a variety of co-variates and functional forms to forecast the returns and prices of metal commodities. Using a large number of models (N large) and a large number of time periods (T large), we apply the techniques put forth by the common-feature literature on forecast combinations. Empirically, we show that models incorporating (short-run) common-cycle restrictions perform better than unrestricted models, with an important role for industrial production as a predictor for metal-price variation.

Bayesian model averaging of load demand forecasts from neural network models

Creating a set of a number of neural network (NN) models in an ensemble and accumulating them can achieve better overview capability as compared to single neural network. Neural network ensembles are designed to provide solutions to particular problems. Many researchers and academicians have adopted this NN ensemble technique, especially in machine learning, and has been applied in various fields of engineering, medicine and information technology. This paper present a robust aggregation methodology for load demand forecasting based on Bayesian Model Averaging of a set of neural network models in an ensemble. This paper estimate a vector of coefficient for individual NN models' forecasts using validation data-set. These coefficients, also known as weights, are equal to posterior probabilities of the models generating the forecasts. These BMA weights are then used in combining forecasts generated from NN models with test data-set. By comparing the Bayesian results with the Simple Averaging method, it was observed that benefits are obtained by utilizing an advanced method like BMA for forecast combinations.

Improving the quality of prediction intervals through optimal aggregation

Neural networks (NNs) are an effective tool to model nonlinear systems. However, their forecasting performance significantly drops in the presence of process uncertainties and disturbances. NN-based prediction intervals (PIs) offer an alternative solution to appropriately quantify uncertainties and disturbances associated with point forecasts. In this paper, an NN ensemble procedure is proposed to construct quality PIs. A recently developed lower-upper bound estimation method is applied to develop NN-based PIs. Then, constructed PIs from the NN ensemble members are combined using a weighted averaging mechanism. Simulated annealing and a genetic algorithm are used to optimally adjust the weights for the aggregation mechanism. The proposed method is examined for three different case studies. Simulation results reveal that the proposed method improves the average PI quality of individual NNs by 22%, 18%, and 78% for the first, second, and third case studies, respectively. The simulation study also demonstrates that a 3%-4% improvement in the quality of PIs can be achieved using the proposed method compared to the simple averaging aggregation method.

Microfounded forecasting

Our focus is on information in expectation surveys that can now be built on thousands (or millions) of respondents on an almost continuous-time basis (big data) and in continuous macroeconomic surveys with a limited number of respondents. We show that, under standard microeconomic and econometric techniques, survey forecasts are an affine function of the conditional expectation of the target variable. This is true whether or not the survey respondent knows the data-generating process (DGP) of the target variable or the econometrician knows the respondents individual loss function. If the econometrician has a mean-squared-error risk function, we show that asymptotically efficient forecasts of the target variable can be built using Hansens (Econometrica, 1982) generalized method of moments in a panel-data context, when N and T diverge or when T diverges with N xed. Sequential asymptotic results are obtained using Phillips and Moon s (Econometrica, 1999) framework. Possible extensions are also discussed.

Forecasting errors, directional accuracy and profitability of currency trading: The case of EUR/USD exchange rate

We provide a comprehensive study of out-of-sample forecasts for the EUR/USD exchange rate based on multivariate macroeconomic models and forecast combinations. We use profit maximization measures based on directional accuracy and trading strategies in addition to standard loss minimization measures. When comparing predictive accuracy and profit measures, data snooping bias free tests are used. The results indicate that forecast combinations, in particular those based on principal components of forecasts, help to improve over benchmark trading strategies, although the excess return per unit of deviation is limited.