Adaptive entropy-based learning with dynamic artificial neural network (original) (raw)
Related papers
A Forecasting Time Series model Based on Entropy and Fuzzy logic
Passer Journal of Basic and Applied Sciences
Electricity Power Consumption Forecasting (EPCF) plays an essential role in global electricity distribution systems that has a significant impact on the operation, control, and planning for the production and distribution of electricity. Due to the complexity, and uncertainty of electricity consumption, especially when the amount of load consumed during different hours is not the same, performing forecasting by using the classical method is inaccurate. To strengthen the efficiency, the time series method that uses a fuzzy approach based on refined entropy is presented in the upcoming article. First, given the specified features, the minimization principle approach of entropy (MPAE) is pursued to define the longitude of each interval in the world of discourse. Secondly, a fuzzy relation matrix of time-invariant is constructed according to the first-order model of fuzzy time series, and the minimum fixed amount of time that the data approach the steady state is obtained using the entropy of the fuzzy set, respectively. Eventually, the forecast results are calculated based on the operation of the maximum combination and the principle of full membership. To show the whole forecasting process, hourly data from July 2022 to September 2022 in Sulaymaniyah / Iraq province is used. Results are compared to the traditional statistical (ARIMA) model, and it indicates that the mean squared error and other criteria of the forecasting error in the entropy based on the fuzzy method are significantly better than the traditional statistical model.
A Method for Estimating the Entropy of Time Series Using Artificial Neural Networks
Entropy
Measuring the predictability and complexity of time series using entropy is essential tool designing and controlling a nonlinear system. However, the existing methods have some drawbacks related to the strong dependence of entropy on the parameters of the methods. To overcome these difficulties, this study proposes a new method for estimating the entropy of a time series using the LogNNet neural network model. The LogNNet reservoir matrix is filled with time series elements according to our algorithm. The accuracy of the classification of images from the MNIST-10 database is considered as the entropy measure and denoted by NNetEn. The novelty of entropy calculation is that the time series is involved in mixing the input information in the reservoir. Greater complexity in the time series leads to a higher classification accuracy and higher NNetEn values. We introduce a new time series characteristic called time series learning inertia that determines the learning rate of the neural n...
A method for estimating the entropy of time series using artificial neural network
ArXiv, 2021
Measuring the predictability and complexity of time series is an essential tool in designing and controlling the nonlinear system. Different entropy measures exist in the literature to analyze the predictability and complexity of time series. However, the existed methods have some drawbacks related to a strong dependence of entropy on the parameters of the methods, as well as on the length and amplitude of the time series. To overcome these difficulties, this study proposes a new method for estimating the entropy of a time series using the LogNNet neural network model. The LogNNet reservoir matrix is filled with the time series elements according to our algorithm. The network is trained on MNIST-10 dataset and the classification accuracy is calculated. The accuracy is considered as the entropy measure and denoted by NNetEn. The novelty of entropy calculation is that the time series is involved in mixing the input information in the reservoir. The greater complexity of the time serie...
Engineering Applications of Artificial Intelligence, 2008
In this paper, next-day hourly forecasts are calculated for the energy price in the electricity production market of Spain. The methodology used to achieve these forecasts is based on artificial neural networks, which have been used successfully in recent years in many forecasting applications. The days to be forecast include working days as well as weekends and holidays, due to the fact that energy price has different behaviours depending on the kind of day to be forecast. Besides, energy price time series are usually composed of too many data, which could be a problem if we are looking for a short period of time to reach an adequate forecast. In this paper, a training method for artificial neural nets is proposed, which is based on making a previous selection for the multilayer perceptron (MLP) training samples, using an ART-type neural network. The MLP is then trained and finally used to calculate forecasts. These forecasts are compared to those obtained from the well-known Box-Jenkins ARIMA forecasting method. Results show that neural nets perform better than ARIMA models, especially for weekends and holidays. Both methodologies calculate more accurate forecasts-in terms of mean absolute percentage error-for working days that for weekends and holidays. Agents involved in the electricity production market, who may need fast forecasts for the price of electricity, would benefit from the results of this study.
Day‐ahead price forecasting of electricity markets by a hybrid intelligent system
European Transactions on Electrical …, 2009
In a competitive electricity market, price forecasts are important for market participants. However, electricity price is a complex signal due to its nonlinearity, nonstationarity, and time variant behavior. In spite of much research in this area, more accurate and robust price forecast methods are still required. In this paper, a combination of a feature selection technique and cascaded neuro-evolutionary algorithm (CNEA) is proposed for this purpose. The feature selection method is an improved version of the mutual information (MI) technique. The CNEA is composed of cascaded forecasters where each forecaster consists of a neural network (NN) and an evolutionary algorithm (EA). An iterative search procedure is also incorporated in our solution strategy to fine-tune the adjustable parameters of both the MI technique and CNEA. The price forecast accuracy of the proposed method is evaluated by means of real data from the Pennsylvania-New Jersey-Maryland (PJM) and Spanish electricity markets. The method is also compared with some of the most recent price forecast techniques.
Forecasting of electricity prices with neural networks
Energy Conversion and Management, 2006
During recent years, the electricity energy market deregulation has led to a new free competition situation in Europe and other countries worldwide. Generators, distributors and qualified clients have some uncertainties about the future evolution of electricity markets. In consequence, feasibility studies of new generation plants, design of new systems and energy management optimization are frequently postponed. The ability of forecasting energy prices, for instance the electricity prices, would be highly appreciated in order to improve the profitability of utility investments.
Short term forecasting of the electricity market of Spain using neural networks
th
Sometimes, having time series that are too long can be an even greater problem, even worse than having series with too few data. From a Neural Networks point of view, it is important to have a set of training samples that is big enough; however, if this set is too big the time required to reach an adequate solution may be too long. In this paper, we propose a training method we have called a selective and continuous method, in which a previous selection for the Multilayer Perceptron (MLP) training samples is made using an ART-type neural network. The MLP is then trained and finally it is used to make forecasts. We tested the effectiveness of the proposed method, making forecasts for the time series called Daily-Market Hourly Price, part of the Electricity Production Market of Spain.
Price forecasting and validation in the Spanish electricity market using forecasts as input data
International Journal of Electrical Power & Energy Systems, 2016
The electricity sector has been subjected to major changes in the last few years. Previously, there existed a regulated system where electric companies could know beforehand the amount of energy each generator would produce, hence basing their largely operational strategy on cost minimization in order to increase their profits. In Spain, from 1988 till 1997, electricity prices were established by the 'Marco Legal Estable'-Stable Legal Framework-, where the Ministry of Industry and Energy acknowledged the existence of certain generation costs related to each type of technology. It was an industrial sector with no actual competition and therefore, with very few controllable risks. In the aftermath of the electricity market liberalization competition and uncertainty arose. Electricity spot prices became highly volatile due to the specific characteristics of electricity as a commodity. Long-term contracts allowed for hedge funds to act against price fluctuation in the electricity market. As a consequence, developing an accurate electricity price forecasting model is an extremely difficult task for electricity market agents. This work aims to propose a methodology to improve the limitations of those methodologies just using historical data to forecast electricity prices. In this manner, and in order to gain access to more recent data, instead of using natural gas prices and electricity load historical data, a regression model to forecast the evolution of natural gas prices, and a model based on artificial neural networks (ANN) to forecast electricity loads, are proposed. The results of these models are used as input for an electricity price forecast model. Finally, and to demonstrate the effectiveness of the proposed methodology, several study cases applied to the Spanish market, using real price data, are presented.
Energies, 2016
This paper presents novel intraday session models for price forecasts (ISMPF models) for hourly price forecasting in the six intraday sessions of the Iberian electricity market (MIBEL) and the analysis of mean absolute percentage errors (MAPEs) obtained with suitable combinations of their input variables in order to find the best ISMPF models. Comparisons of errors from different ISMPF models identified the most important variables for forecasting purposes. Similar analyses were applied to determine the best daily session models for price forecasts (DSMPF models) for the day-ahead price forecasting in the daily session of the MIBEL, considering as input variables extensive hourly time series records of recent prices, power demands and power generations in the previous day, forecasts of demand, wind power generation and weather for the day-ahead, and chronological variables. ISMPF models include the input variables of DSMPF models as well as the daily session prices and prices of preceding intraday sessions. The best ISMPF models achieved lower MAPEs for most of the intraday sessions compared to the error of the best DSMPF model; furthermore, such DSMPF error was very close to the lowest limit error for the daily session. The best ISMPF models can be useful for MIBEL agents of the electricity intraday market and the electric energy industry.
The effects of pre-processing methods on forecasting improvement of artificial neural networks
Aus. J. Basic. Appl. …, 2011
Along other methods, Intelligent Methods can be used in order to model the trend of changes of a certain variable. These methods require data to be preprocessed before being used in the forecasting process. Generally, the preprocessing step includes omitting outliers, assessment of the missing data, data smoothing, etc. In this paper, the effect of various smoothing methods on the final forecasted results is studied. Furthermore, data from the electricity consumption in Iran over the past 20 years were used as actual data. After being smoothed, these data are then incorporated into an Artificial Neural Network in order to forecast the electrical consumption. The comparisons between several Seasonal Decomposition, including Seasonal Adjustment Series (SAS) and Seasonal Trend Cycle (STC), Exponential Smoothing (Simple, Linear, Holt and Winter) and Box-Jenkins (Moving Average, Auto Regression, and Auto Regression Integrated Moving Average) methods show the superiority of SAS in Decomposition categorization over other methods. The structure of this study may be used for other data sets for improvement of data pre-processing. Fig. 1: The electricity consumption trend in a period of 132 months.