Artificial Neural Network Forecasting Models Research Articles

ARTIFICIAL NEURAL NETWORK BASED DEMAND FORECASTING INTEGRATED WITH FEDERAL FUNDS RATE

Adverse effects of inaccurate demand forecasts; stockouts, overstocks, customer loss have led academia and the business world towards accurate demand forecasting methods. Artificial Neural Network (ANN) is capable of highly accurate forecasts integrated with many variables. The use of Price and Promotion variables have increased the accuracy while the addition of other relevant variables would decrease the occurrences of errors. The use of the Federal Funds Rate as an additional macroeconomic variable to ANN forecasting models has been discussed in this research by the means of the accuracy measuring method: Average Relative Mean Absolute Error.

A Sustainable Early Warning System Using Rolling Forecasts Based on ANN and Golden Ratio Optimization Methods to Accurately Predict Real-Time Water Levels and Flash Flood.

Remote monitoring sensor systems play a significant role in the evaluation and minimization of natural disasters and risk. This article presents a sustainable and real-time early warning system of sensors employed in flash flood prediction by using a rolling forecast model based on Artificial Neural Network (ANN) and Golden Ratio Optimization (GROM) methods. This Early Flood Warning System (EFWS) aims to support decision makers by providing reliable and accurate information and warning about any possible flood events within an efficient lead-time to reduce any damages due to flash floods. In this work, to improve the performance of the EFWS, an ANN forecast model based on a new optimization method, GROM, is developed and compared to the traditional ANN model. Furthermore, due to the lack of literature regarding the optimal ANN structural model for forecasting the flash flood, this paper is one of the first extensive investigations into the impact of using different exogenous variables and parameters on the ANN structure. The effect of using a rolling forecast model compared to fixed model on the accuracy of the forecasts is investigated as well. The results indicate that the rolling ANN forecast model based on GROM successfully improved the model accuracy by 40% compared to the traditional ANN model and by 93.5% compared to the fixed forecast model.

Long-Term Rainfall Information Forecast by Utilizing Constrained Amount of Observation through Artificial Neural Network Approach

Estimating models are becoming increasingly crucial in highlighting the nonlinear connections of the massive level of rough information and chaotic components. The study demonstrates a modern approach utilizing a created artificial neural network (ANN) method that may be an alternative strategy to conventional factual procedures for advancing rainfall estimate execution. A case study was presented for Turkey’s Düzce and Bolu neighboring territories located on the Black Sea’s southern coast. This study’s primary aim is to create an ANN model unique in the field to generate satisfactory results even with limited data. The proposed technique is being used to estimate rainfall and make predictions regarding future precipitation. Bolu daily average rainfall by month data and a limited number of Düzce rainfall data were used. Missing forecasts and potential rainfall projections will be examined in the fundamental research. This research further focuses on ANN computational concepts and develops a neural network for rainfall time series forecasting. The emphasis of this study was a feed-forward backpropagation network. The Levenberg–Marquardt algorithm (LMA) was implemented for training a two-layer feed-forward ANN for the missing rainfall data prediction part of this research. The inaccessible rainfall parameters for Düzce were determined for the years 1995 to 2009. From 2010 to 2020, a two-layer feed-forward ANN was trained using the gradient descent algorithm to forecast daily average rainfall data by month. The findings reported in this study guide researchers interested in implementing the ANN forecast model for an extended period of missing rainfall data.

Medium, short and very short-term prognosis of load demand for the Greek Island of Tilos using artificial neural networks and human thermal comfort-discomfort biometeorological data

The objective of the present work is the medium, short and very short-term prognosis of load demand (LD) for the small-scale island of Tilos in Greece. For this purpose, Artificial Neural Network (ANNs) models were developed to forecast the LD of Tilos for different prediction horizons and time intervals, these covering the cases of 24 h ahead in hourly intervals (medium term prognosis), 2 h ahead in 10-min intervals (short term prognosis) and 10-min ahead in 1-min intervals (very short term prognosis). At the same time, stochastic/persistence autoregressive (AR) models were also developed and compared with the respective ANN models with regards to the LD prediction results obtained.For the training of the developed ANNs, meteorological data covering the period 2015–2017 were used, which had been recorded in 1-min intervals by two meteorological masts installed on the island Tilos. Furthermore, the biometeorological human thermal comfort-discomfort index, known as the cooling power index (CP), was also estimated and introduced in the training procedure of the forecasting models, while, for the evaluation of both AR and ANN forecasting models, well established statistical evaluation indices were applied.To this end, results show that in all cases covered, i.e. for both medium and short-term prognoses, the developed ANN forecasting models present a remarkable ability to predict the local LD of the island with high accuracy, enabling in this way the development of advanced energy management tools for both end-users and the system operators.

Hybrid intra-hour DNI forecasts with sky image processing enhanced by stochastic learning

We propose novel smart forecasting models for Direct Normal Irradiance (DNI) that combine sky image processing with Artificial Neural Network (ANN) optimization schemes. The forecasting models, which were developed for over 6months of intra-minute imaging and irradiance measurements, are used to predict 1min average DNI for specific time horizons of 5 and 10min. We discuss optimal models for low and high DNI variability seasons. The different methods used to develop these season-specific models consist of sky image processing, deterministic and ANN forecasting models, a genetic algorithm (GA) overseeing model optimization and two alternative methods for training and validation. The validation process is carried over by the Cross Validation Method (CVM) and by a randomized training and validation set method (RTM). The forecast performance for each solar variability season is evaluated, and the models with the best forecasting skill for each season are selected to build a hybrid model that exhibits optimal performance for all seasons. An independent testing set is used to assess the performance of all forecasting models. Performance is assessed in terms of common error statistics (mean bias and root mean square error), but also in terms of forecasting skill over persistence. The hybrid forecast models proposed in this work achieve statistically robust forecasting skills in excess of 20% over persistence for both 5 and 10min ahead forecasts, respectively.

Artificial Neural Network–Based Drought Forecasting Using a Nonlinear Aggregated Drought Index

AbstractDrought forecasting plays an important role in the planning and management of water resources systems, especially during dry climatic periods. In this study, a nonlinear aggregated drought index (NADI) was developed first to classify the drought condition of a catchment considering all significant hydrometeorological variables that have effects on droughts. An artificial neural network (ANN)—based drought forecasting approach was then developed by using the time series of the NADI to forecast NADI values. In forecasting future drought conditions, the NADI produces the overall dryness within the system as compared to the traditional forecasting of rainfall deficiency, which considers only the meteorological droughts. Two ANN forecasting models, namely a recursive multistep neural network (RMSNN) and a direct multistep neural network (DMSNN), were developed in this study. Overall, these models were capable of forecasting drought conditions well for up to 6 months of future forecasts, which were stat...

Assessment of Arias Intensity of historical earthquakes using modified Mercalli intensities and artificial neural networks

Abstract. This paper presents the development of a non-parametric forecast model based on artificial neural networks for the direct assessment of Arias Intensity corresponding to a historic earthquake using seismic intensity data. The neural models allow complex and nonlinear behaviour to be tracked. Application of this methodology on earthquakes with known instrumental data from Greece, showed that the artificial neural network forecast model have excellent data synthesis capability.