Conventional Forecasting Models Research Articles

Research and Application of a Novel Combined Model Based on Multiobjective Optimization for Multistep-Ahead Electric Load Forecasting

Accurate forecasting of electric loads has a great impact on actual power generation, power distribution, and tariff pricing. Therefore, in recent years, scholars all over the world have been proposing more forecasting models aimed at improving forecasting performance; however, many of them are conventional forecasting models which do not take the limitations of individual predicting models or data preprocessing into account, leading to poor forecasting accuracy. In this study, to overcome these drawbacks, a novel model combining a data preprocessing technique, forecasting algorithms and an advanced optimization algorithm is developed. Thirty-minute electrical load data from power stations in New South Wales and Queensland, Australia, are used as the testing data to estimate our proposed model’s effectiveness. From experimental results, our proposed combined model shows absolute superiority in both forecasting accuracy and forecasting stability compared with other conventional forecasting models.

Two hybrid data-driven models for modeling water-air temperature relationship in rivers.

River water temperature (RWT) forecasting is important for the management of stream ecology. In this paper, a new method based on coupling of wavelet transformation (WT) and artificial intelligence (AI) techniques, including multilayer perceptron neural network (MLPNN) and adaptive neural-fuzzy inference system (ANFIS) for RWT prediction is proposed. The performances of the hybrid models are compared with regular MLPNN and ANFIS models and multiple linear regression (MLR) models for RWT forecasting in two river stations in the Drava River, Croatia. Model performance was evaluated using the coefficient of correlation (R), the Willmott index of agreement (d), the root mean squared error (RMSE), and the mean absolute error (MAE). Results indicate that the combination of WT and AI models (WTMLPNN and WTANFIS) yield better models than the conventional forecasting models for RWT simulation for both regular periods and heatwave events. The MLPNN and ANFIS models outperform the MLR models for RWT simulation for the studied river stations. RMSE values of WTMLPNN2 and WTANFIS2 models range from 1.127 to 1.286°C, and 1.216 to 1.491°C for the Botovo and Donji Miholjac stations respectively. Additionally, modeling results further confirm the importance of the day of year (DOY) on the thermal dynamics of the river. The results of this study indicate the potential of coupling of WT and MLPNN, ANFIS models in forecasting RWT.

Day-Ahead System Marginal Price Forecasting Using Artificial Neural Network and Similar-Days Information

Day-ahead system marginal price (SMP) forecasting constitutes essential information in the competitive energy market. Hence, this paper presents the development of a day-ahead SMP forecasting model via implementing an artificial neural network (ANN) algorithm. The accuracy of the ANN-based model is improved by including long-term historical data in addition to short-term historical data and by applying the k-fold cross-validation optimization algorithm. The selection of the short-term type input variable applies the Pearson correlation coefficient. Whereas the long-term type input variable is selected by applying the discrete Frechet distance in conjunction with the information related to the season and type of the day to find the Similar-Days Index. In order to verify the model, the forecasted load and actual SMP for 15 years of historical data are used. The results indicate that the proposed model can forecast SMP with higher accuracy than the conventional forecasting model.

A Hybrid Short-term Traffic Flow Forecasting Method Based on Neural Networks Combined with K-Nearest Neighbor

It is critical to implement accurate short-term traffic forecasting in traffic management and control applications. This paper proposes a hybrid forecasting method based on neural networks combined with the K-nearest neighbor (K-NN) method for short-term traffic flow forecasting. The procedure of training a neural network model using existing traffic input-output data, i.e., training data, is indispensable for fine-tuning the prediction model. Based on this point, the K-NN method was employed to reconstruct the training data for neural network models while considering the similarity of traffic flow patterns. This was done through collecting the specific state vectors that were closest to the current state vectors from the historical database to enhance the relationship between the inputs and outputs for the neural network models. In this study, we selected four different neural network models, i.e., back-propagation (BP) neural network, radial basis function (RBF) neural network, generalized regression (GR) neural network, and Elman neural network, all of which have been widely applied for short-term traffic forecasting. Using real world traffic data, the experimental results primarily show that the BP and GR neural networks combined with the K-NN method have better prediction performance, and both are sensitive to the size of the training data. Secondly, the forecast accuracies of the RBF and Elman neural networks combined with the K-NN method both remain fairly stable with the increasing size of the training data. In summary, the proposed hybrid forecasting approach outperforms the conventional forecasting models, facilitating the implementation of short-term traffic forecasting in traffic management and control applications.

A novel combined model based on advanced optimization algorithm for short-term wind speed forecasting

Short-term wind speed forecasting has a significant influence on enhancing the operation efficiency and increasing the economic benefits of wind power generation systems. A substantial number of wind speed forecasting models, which are aimed at improving the forecasting performance, have been proposed. However, some conventional forecasting models do not consider the necessity and importance of data preprocessing. Moreover, they neglect the limitations of individual forecasting models, leading to poor forecasting accuracy. In this study, a novel model combining a data preprocessing technique, forecasting algorithms, an advanced optimization algorithm, and no negative constraint theory is developed. This combined model successfully overcomes some limitations of the individual forecasting models and effectively improves the forecasting accuracy. To estimate the effectiveness of the proposed combined model, 10-min wind speed data from the wind farm in Peng Lai, China are used as case studies. The experiment results demonstrate that the developed combined model is definitely superior compared to all other conventional models. Furthermore, it can be used as an effective technique for smart grid planning.

An adaptive middle and long-term runoff forecast model using EEMD-ANN hybrid approach

It remains a challenge to obtain an accurate middle and long-term runoff forecast, especially in flood seasons. The forecast performance can be improved using ensemble empirical mode decomposition (EEMD) to produce cleaner signals as model inputs. In many EEMD based forecast models, the entire time series are decomposed into several sub-series, and each sub-series is divided into calibration and validation datasets and forecasted by some common models, such as artificial neural network (ANN), and finally an ensemble forecast is obtained by summing the forecasted results of each sub-series. In such a decomposition-ensemble framework, some future information is used, and thus it is a hindcast experiment. Attempts have also been made to propose a real forecast experiment, which, however, often cannot be able to adapt to the non-stationary changes in runoff due to the lack of adaptive ability. Therefore, this study tries to improve the decomposition-ensemble framework and propose an adaptive middle and long-term runoff forecast model especially for flood seasons. Unlike a hindcast experiment, it does not use any future information; and unlike conventional forecast experiments, its decomposition and forecast models are adjusted adaptively as long as new runoff information is added. EEMD is used to decompose the original time series and ANN is used to forecast each sub-series, hence the name an adaptive EEMD-ANN (AEEMD-ANN) model. This model is applied to forecast the 1-month ahead streamflow of three stations in China, and the results show that the AEEMD-ANN model can improve forecast accuracy in flood seasons, but it is not as good as ANN, adaptive neuro-fuzzy inference system (ANFIS), support vector machine (SVM) and seasonal first-order autoregressive (SAR (1)) models in dry seasons. While the conventional forecast models, especially the SAR (1) models, are more suitable in dry season. Therefore, it is recommended to use SAR (1) model in dry season and AEEMD-ANN model in flood season to forecast the monthly runoff in Yangtze River Basin.

Intention-based electoral forecasting: limits of accuracy of conventional model and its development perspectives with regard to emotional factors

Over the recent years there have been disruptions of the functioning of the conventional model of electoral forecasting. The reason for the disruptions remains unclear, and field-specific studies following the most dramatic mistakes are focused on common accuracy-related issues (samples’ bias, socially desirable responding, etc.). The present article argues that inaccuracy take place as the conventional model fails to take into account emotional factors of voting behavior. Using empirical evidence of the 2016 State Duma and the 2018 presidential elections the author examines the applicability of Rosenberg’s consistency theory to the voting behavior and the influence of an additional filter (affective attitudes of voters towards candidates) on the accuracy of the results. The findings suggest that mismatching attitude towards different electoral alternatives is well fixed and regular. In this respect, adding the factor of affective attitudes as an additional filter in the forecasting model can continuously increase the forecasting accuracy. In full accordance with the proposed model, this phenomenon is well observed in the case of higher political uncertainty and a relative mismatch in attitudes (2016) and least observed in the case of lower political uncertainty and a relative attitude-behavior consistency (2018). The importance of the problem formulated by the author arises from a high probability of significant political uncertainty in the 2021 and 2024 elections. In terms of theory and methodology, the obtained results raise a question about possible improvements essential for the conventional forecasting model.

Deterministic and probabilistic forecasting of photovoltaic power based on deep convolutional neural network

The penetration of photovoltaic (PV) energy into modern electric power and energy systems has been gradually increased in recent years due to its benefits of being abundant, inexhaustible and clean. In order to reduce the negative impacts of PV energy on electric power and energy systems, advanced forecasting approach with high-accuracy is a pressing need. Aimed at this, a novel hybrid method for deterministic PV power forecasting based on wavelet transform (WT) and deep convolutional neural network (DCNN) is firstly proposed in this paper. WT is used to decompose the original signal into several frequency series. Each frequency has better outlines and behaviors. DCNN is employed to extract the nonlinear features and invariant structures exhibited in each frequency. Then, a probabilistic PV power forecasting model that combines the proposed deterministic method and spine quantile regression (QR) is originally developed to statistically evaluate the probabilistic information in PV power data. The proposed deterministic and probabilistic forecasting methods are applied to real PV data series collected from PV farms in Belgium. Numerical results presented in the case studies demonstrate that the proposed methods exhibit the ability of improving forecasting accuracies in terms of seasons and various prediction horizons, when compared to conventional forecasting models.

Forecasting Construction Cost Index Using Interrupted Time-Series

The Construction Cost Index (CCI) provided by the Engineering News-Record (ENR) is frequently used in predicting the cost of construction projects because it reflects a comprehensive trend in the construction costs. In previous studies, the CCI was forecasted using the time-series analysis methods. However, the effects of specific intervening factors such as economic recession and policy changes on the CCI data were neglected, thus compromising the reliability and accuracy of the forecast models. In this study, an interrupted time-series forecasting model was developed wherein the economic recession of 2008 was reflected in the forecasting model, which is an outlier identified to have significant impact on the CCI. The forecast result obtained using the interrupted timeseries forecasting model was better than that using the conventional forecast models such as the ARIMA and Holt–Winters exponential-smoothing models. The accurately forecasted CCI using the presented model will help in budget and bid planning as well as assessing the risk of future businesses.

LSTM network: a deep learning approach for short‐term traffic forecast

Short-term traffic forecast is one of the essential issues in intelligent transportation system. Accurate forecast result enables commuters make appropriate travel modes, travel routes, and departure time, which is meaningful in traffic management. To promote the forecast accuracy, a feasible way is to develop a more effective approach for traffic data analysis. The availability of abundant traffic data and computation power emerge in recent years, which motivates us to improve the accuracy of short-term traffic forecast via deep learning approaches. A novel traffic forecast model based on long short-term memory (LSTM) network is proposed. Different from conventional forecast models, the proposed LSTM network considers temporal-spatial correlation in traffic system via a two-dimensional network which is composed of many memory units. A comparison with other representative forecast models validates that the proposed LSTM network can achieve a better performance.

A novel forecasting model based on a hybrid processing strategy and an optimized local linear fuzzy neural network to make wind power forecasting: A case study of wind farms in China

As a crucial issue in the wind power industry, it is a tough and challenging task to predict the wind power accurately because of its nonlinearity, non-stationary and chaos. In this paper, we propose a novel hybrid model, which combines an integrated processing strategy and an optimized local linear fuzzy neural network, to forecast the wind power. First, discrete wavelet transform and singular spectrum analysis are used to filter out the noises and extract the trends from original wind power series, respectively. Then, the novel no-negative-constraint-combination theory together with the CS algorithm are adopted to integrate these two subseries obtained from the first step to retain strengths of each method. Based on the phase space reconstruction model, we could determine the most proper structure of the input sets and the output sets. Next, the local linear fuzzy neural network, with the initial rule consequent parameters optimized by the seeker optimization algorithm, is utilized to make wind power forecasts for a selected number of forward time steps. The numerical results from two experiments demonstrate that the proposed hybrid model is an effective approach to predict wind power, and the accuracy of prediction is highly improved compared with conventional forecasting models.

RAINFALL-INDUCED MECHANISM BASED EARLY WARNING MODEL FOR SLOPES OF DUALISTIC LAYERS IN WEATHERED GRANITIC AREA

Rainfall-induced landslides are clustery, recorded by Chinese historic documents. Study of early-warning is a significant way to prevent and reduce damages. Conventional forecasting model is based on statistics of landslides, which is short of inducement mechanism and mechanical mechanism, with limited dependability and precision of forecasting. The typical landslide in weathered granitic crust area is studied in this paper. Based on simplified geological model and saturated-unsaturated seepage theory, the process of rainfall infiltration and instability mechanism is studied, then the warning criterion of typical slope is established. (1) Typical dual-structure slopes in weathered granite crust area are soil-like slopes and covering soil is thick. They are divided into two parts on profile. The upper part is collusive clay with loose soil and high permeability, while the lower part is residual clay with compactness soil and lower permeability. (2) Rainfall infiltration under several conditions are simulated. Taking rainfall intensity of 50mm·d-1 for example, within 30 hours, most rain is in upper collusive clay, causing front edge of slope saturated and sliding zone hydropsy. From 40 to 50 hours, infiltration along with tension cracks around rear edge of slope is faster, and water in cracks may generate hydrostatic increasing weight of soil and sliding power. Meanwhile toe of slope saturates and breaks firstly. Then landslide occurs. (3) Porewater pressure increases continually from monitoring in three parts of slope(front, central and rear of slope). Namely slope soil transforms unsaturated to saturated. The toe of slope gets saturated first, central part under continual infiltration, and upper part saturation cause sliding zone deformation and sliding through hydrostatic pressure. (4) Durations needed when landslides occur under different rainfall conditions are:13 days with moderate rain(10mm·d-1), 5 days with heavy rain(25mm·d-1), 2.2 days with rainstorm(50mm·d-1), and 1.1 days with extraordinary rainstorm(100mm·d-1). Duration of hysteresis is about 5 hours when rainstorm. Finally, critical rainfall criterion(I-D curve) of this kind of slopes is established.

Forecasting Local Inflation with Global Inflation: When Economic Theory Meets the Facts

This paper provides both theoretical insight as well as empirical evidence in support of the view that inflation is largely a global phenomenon. First, we show that inflation across countries incorporates a significant common factor captured by global inflation. Second, we show that in theory a role for global inflation in local inflation dynamics emerges over the business cycle even without common shocks, and under flexible exchange rates and complete international asset markets. Third, we identify a strong correction mechanism that brings local inflation rates back in line with global inflation which explains the relative success of inflation forecasting models based on global inflation (e.g., Ciccarelli and Mojon (2010)). Fourth, we argue that the workhorse New Open Economy Macro (NOEM) model of Martinez-Garcia and Wynne (2010) can be approximated by a finite-order VAR and estimated using Bayesian techniques to forecast domestic inflation incorporating all relevant linkages with the rest of the world. This NOEM-BVAR provides a tractable model of inflation determination that can be tested empirically in forecasting. Finally, we use pseudo-out-of-sample forecasts to assess the NOEM-BVAR at different horizons (1 to 8 quarters ahead) across 17 OECD countries using quarterly data over the period 1980Q1-2014Q4. In general, we find that the NOEM-BVAR model produces a lower root mean squared prediction error (RMSPE) than its competitors --- which include most conventional forecasting models based on domestic factors and also the recent models based on global inflation. In a number of cases, the gains in smaller RMSPEs are statistically significant. The NOEM-BVAR model is also accurate in predicting the direction of change for inflation, and often better than its competitors along this dimension too.

Forecasting Local Inflation with Global Inflation: When Economic Theory Meets the Facts

This paper provides both theoretical insight as well as empirical evidence in support of the view that inflation is largely a global phenomenon. First, we show that inflation across countries incorporates a significant common factor captured by global inflation. Second, we show that in theory a role for global inflation in local inflation dynamics emerges over the business cycle even without common shocks, and under flexible exchange rates and complete international asset markets. Third, we identify a strong correction mechanism that brings local inflation rates back in line with global inflation which explains the relative success of inflation forecasting models based on global inflation (e.g., Ciccarelli and Mojon (2010). Fourth, we argue that the workhorse New Open Economy Macro (NOEM) model of Martinez-Garcia and Wynne (2010) can be approximated by a finiteorder VAR and estimated using Bayesian techniques to forecast domestic inflation incorporating all relevant linkages with the rest of the world. This NOEM-BVAR provides a tractable model of inflation determination that can be tested empirically in forecasting. Finally, we use pseudo-out-of-sample forecasts to assess the NOEM-BVAR at different horizons (1 to 8 quarters ahead) across 17 OECD countries using quarterly data over the period 1980Q1-2014Q4. In general, we find that the NOEM-BVAR model produces a lower root mean squared prediction error (RMSPE) than its competitors—which include most conventional forecasting models based on domestic factors and also the recent models based on global inflation. In a number of cases, the gains in smaller RMSPEs are statistically significant. The NOEM-BVAR model is also accurate in predicting the direction of change for inflation, and often better than its competitors along this dimension too.

Forecasting Local Inflation with Global Inflation: When Economic Theory Meets the Facts

This paper provides both theoretical insight as well as empirical evidence in support of the view that inflation is largely a global phenomenon. First, we show that inflation across countries incorporates a significant common factor captured by global inflation. Second, we show that in theory a role for global inflation in local inflation dynamics emerges over the business cycle even without common shocks, and under flexible exchange rates and complete international asset markets. Third, we identify a strong \error correction mechanism that brings local inflation rates back in line with global inflation which explains the relative success of inflation forecasting models based on global inflation (e.g., Ciccarelli and Mojon (2010). Fourth, we argue that the workhorse New Open Economy Macro (NOEM) model of Martnez-Garca and Wynne (2010) can be approximated by a finiteorder VAR and estimated using Bayesian techniques to forecast domestic inflation incorporating all relevant linkages with the rest of the world. This NOEM-BVAR provides a tractable model of inflation determination that can be tested empirically in forecasting. Finally, we use pseudo-out-of-sample forecasts to assess the NOEM-BVAR at different horizons (1 to 8 quarters ahead) across 17 OECD countries using quarterly data over the period 1980Q1-2014Q4. In general, we find that the NOEM-BVAR model produces a lower root mean squared prediction error (RMSPE) than its competitors?which include most conventional forecasting models based on domestic factors and also the recent models based on global inflation. In a number of cases, the gains in smaller RMSPEs are statistically significant. The NOEM-BVAR model is also accurate in predicting the direction of change for inflation, and often better than its competitors along this dimension too.

A panel vector error correction approach to forecasting demand in regional construction markets

Reliable forecasting as to the level of aggregate demand for construction is of vital importance to developers, builders and policymakers. Previous construction demand forecasting studies mainly focused on temporal estimating using national aggregate data. The construction market can be better represented by a group of interconnected regions or local markets rather than a national aggregate, and yet regional forecasting techniques have rarely been applied. Furthermore, limited research has applied regional variations in construction markets to construction demand modelling and forecasting. A new comprehensive method is used, a panel vector error correction approach, to forecast regional construction demand using Australia’s state-level data. The links between regional construction demand and general economic indicators are investigated by panel cointegration and causality analysis. The empirical results suggest that both long-run and causal links are found between regional construction demand and construction price, state income, population, unemployment rates and interest rates. The panel vector error correction model can provide reliable and robust forecasting with less than 10% of the mean absolute percentage error for a medium-term trend of regional construction demand and outperforms the conventional forecasting models (panel multiple regression and time series multiple regression model). The key macroeconomic factors of construction demand variations across regions in Australia are also presented. The findings and robust econometric techniques used are valuable to construction economists in examining future construction markets at a regional level.

Daily water level forecasting using wavelet decomposition and artificial intelligence techniques

Summary Reliable water level forecasting for reservoir inflow is essential for reservoir operation. The objective of this paper is to develop and apply two hybrid models for daily water level forecasting and investigate their accuracy. These two hybrid models are wavelet-based artificial neural network (WANN) and wavelet-based adaptive neuro-fuzzy inference system (WANFIS). Wavelet decomposition is employed to decompose an input time series into approximation and detail components. The decomposed time series are used as inputs to artificial neural networks (ANN) and adaptive neuro-fuzzy inference system (ANFIS) for WANN and WANFIS models, respectively. Based on statistical performance indexes, the WANN and WANFIS models are found to produce better efficiency than the ANN and ANFIS models. WANFIS7-sym10 yields the best performance among all other models. It is found that wavelet decomposition improves the accuracy of ANN and ANFIS. This study evaluates the accuracy of the WANN and WANFIS models for different mother wavelets, including Daubechies, Symmlet and Coiflet wavelets. It is found that the model performance is dependent on input sets and mother wavelets, and the wavelet decomposition using mother wavelet, db10, can further improve the efficiency of ANN and ANFIS models. Results obtained from this study indicate that the conjunction of wavelet decomposition and artificial intelligence models can be a useful tool for accurate forecasting daily water level and can yield better efficiency than the conventional forecasting models.

Tourism Demand Forecasting Based on a Neuro-Fuzzy Model

Tourism in Greece plays a major role in the country's economy and an accurate forecasting model for tourism demand is a useful tool, which could affect decision making and planning for the future. This paper answers some questions such as: how did the forecasting techniques evolve over the years, how precise can they be, and in what way can they be used in assessing the demand for tourism? An Adaptive Neuro-Fuzzy Inference System (ANFIS) has been used in making the forecasts. The data used as input for the forecasting models relates to monthly time-series tourist arrivals by air, train, sea and road into Greece from January 1996 until September 2011. 80% of the data has been used to train the forecasting models and the rest to evaluate the models. The performance of the model is achieved by the calculation of some well known statistical errors. The accuracy of the ANFIS model is further compared with two conventional forecasting models: the autoregressive (AR) and autoregressive moving average (ARMA) time-series models. The results were satisfactory even if the collected data were not pleasing enough. The ANFIS performed further compared to the other time-series models. In conclusion, the accuracy of the ANFIS model forecast proved its great importance in tourism demand forecasting.

Relativistic electron flux forecast at geostationary orbit using Kalman filter based on multivariate autoregressive model

The relativistic electron population at MeV energy in the Van Allen radiation belts at geostationary orbit largely varies in association with solar wind disturbances. To provide alerts of possible satellite malfunctions due to deep‐dielectric charging during relativistic electron enhancements, the National Institute of Information and Communications Technology, Japan, developed an algorithm to forecast daily >2 MeV electron flux variations at geostationary orbit using a multivariate autoregressive model. We examined model accuracy by using solar wind speed, north‐south component of the magnetic field, and dynamic pressure by inputting them as explanatory variates. The results showed that a combination of all three variates was most effective in reducing the prediction error. We focus here on the four‐variate autoregressive model and handle it using the Kalman filter. The time evolution of the forecast is given by the conditional normal distribution: the peak value of forecast probability and the error range. The error range estimation is useful for users who utilize forecasts for operation of the satellites. We investigated the prediction efficiency of +1 day forecasts by evaluating forecast and observation data for a whole solar cycle (1999–2008) every 2 years. The prediction efficiency maintained at more than 69% throughout the solar cycle, although it depends on the phase of the cycle. Comparisons of the prediction efficiencies revealed that our model exhibited the best performance of conventional forecast models, particularly in solar active periods.

Using dynamic factor models to forecast Canadian inflation: the role of US variables1

This article evaluates the forecasting performance of dynamic factor models for Canadian inflation. We find that factor models are as good as more conventional forecasting models, while a model estimated using only US data, is at least as useful as a model that incorporates Canadian data. This suggests that the United States is a source of data that could be beneficial to its trading partners for forecasting purposes.