The Application of LSM for Long-Term Forecasting of Specific Fuel Consumption by the Energy System of Uzbekistan

Abstract

The article considers the possibility of using the least squares method (LSM) for long-term forecasting of the parameters of the regime of electric power systems. There is presented least squares method for predicting the parameters of the regime of electric power systems. It is shown that, based on the least-squares method, it is possible to obtain prognostic equations, aswell as coefficients of approximating functions necessary for the formation of these equations. The results of the analysis of the comparison of linear, hyperbolic, logarithmic, exponential and quadratic functions on the use of LSMs to predict specific fuel consumption are presented. Analytical studies are based on statistical data on specific fuel consumption for the period 1990-2016 years by the power system of Uzbekistan. There is shown that the statistical data was divided into training and control samples, when performing an analysis of comparisons of algebraic functions. The training sample, which based on prediction equations are obtained using algebraic functions of various types. The criterion of the least squares method, which is according for using the statistical data of the control sample in the obtained prognostic functions, the standard deviations are found. In the end, there has drawn conclusions, based on the obtained results.

Application of computer simulation results and machine learning in analysis of microwave radiothermometry data

“…It can be assumed that the efficiency measure in the first groups decreased in comparison with the works results [8] and [9] due to a decrease in the feature space. Namely, the thermoasymmetry indicators, one of the most effective feature space elements, were not used.…”

“…On the other hand, in recent years, the possibilities of using machine learning algorithms in the formulation and a diagnostic solution based on microwave radiothermometry data substantiation have been actively studied [2,4,8,9]. It was quickly established that an attempt to make a diagnosis using artificial intelligence methods based solely on temperature data does not provide the required sensitivity and specificity.…”

“…However, here, too, specific input signal preprocessing based on the corresponding feature spaces [8] made it possible to significantly increase the accuracy. Namely, the combined modeling functions and the anamnesis results usage in the first layer, proposed in [9], made it possible to increase the classifier efficiency by more than 10 percent compared to using only temperature values in the input layer [2]. The sensitivity and specificity in the artificial neural network model proposed in [9] reached values of 0.8 -0.85.…”