Wyniki wyszukiwania - BazTech

1

Bagging and boosting techniques in prediction of particulate matters

Triana D., Osowski S.

Bulletin of the Polish Academy of Sciences. Technical Sciences

|

2020

|

Vol. 68, nr 5

1207--1215

EN

The paper presents new ensemble solutions, which can forecast the average level of particulate matters PM10 and PM2.5 with increased accuracy. The proposed network is composed of weak predictors integrated into a final expert system. The members of the ensemble are built based on deep multilayer perceptron and decision tree and use bagging and boosting principle in elaborating common decisions. The numerical experiments have been carried out for prediction of daily average pollution of PM10 and PM2.5 for the next day. The results of experiments have shown, that bagging and boosting ensembles employing these weak predictors improve greatly the quality of results. The mean absolute errors have been reduced by more than 30% in the case of PM10 and 20% in the case of PM2.5 in comparison to individually acting predictors.

2

Local dynamic integration of ensemble in prediction of time series

Osowski S., Siwek K.

Bulletin of the Polish Academy of Sciences. Technical Sciences

|

2019

|

Vol. 67, nr 3

517--525

EN

The paper presents local dynamic approach to integration of an ensemble of predictors. The classical fusing of many predictor results takes into account all units and takes the weighted average of the results of all units forming the ensemble. This paper proposes different approach. The prediction of time series for the next day is done here by only one member of an ensemble, which was the best in the learning stage for the input vector, closest to the input data actually applied. Thanks to such arrangement we avoid the situation in which the worst unit reduces the accuracy of the whole ensemble. This way we obtain an increased level of statistical forecasting accuracy, since each task is performed by the best suited predictor. Moreover, such arrangement of integration allows for using units of very different quality without decreasing the quality of final prediction. The numerical experiments performed for forecasting the next input, the average PM10 pollution and forecasting the 24-element vector of hourly load of the power system have confirmed the superiority of the presented approach. All quality measures of forecast have been significantly improved.

3

Local dynamic integration of ensemble of predictors in load forecasting

Ciechulski T., Osowski S.

Przegląd Elektrotechniczny

|

2017

|

R. 93, nr 10

57--59

EN

The paper shows the new approach to integration of an ensemble of neural predictors in load forecasting. In opposite to classic integration method built upon weighted averaging of every single predictor results this integration method uses only the results of one predictor which was the best on the input data of the learning vectors from the past, which were closest to the actual excitation. Thanks to this the result of ensemble is never worse than the best unit in ensemble. The results of 24-hour ahead prediction of the daily load in small power system have confirmed the efficiency of the proposed solution.

PL

Artykuł przedstawia nowe podejście do integracji zespołu predyktorów neuronowych w zadaniu prognozowania godzinnych obciążeń dobowych z wyprzedzeniem 24-godzinnym. W metodyce tej do predykcji używany jest tylko jeden – najlepszy predyktor dla analizowanej doby. Konkretny wektor obciążeń z danych uczących wraz z najbardziej dokładną odpowiadającą mu siecią neuronową wyłonioną w trybie uczenia wybierany jest na podstawie najmniejszej odległości euklidesowej badanego wektora w trybie testującym. Wyniki badań numerycznych potwierdzają wyższość prezentowanej metody nad rozwiązaniami klasycznymi predykcji.

4

Ensemble neural network approach for accurate load forecasting in a power system

Siwek K., Osowski S., Szupiluk R.

International Journal of Applied Mathematics and Computer Science

|

2009

|

Vol. 19, no 2

303-315

EN

The paper presents an improved method for 1-24 hours load forecasting in the power system, integrating and combining different neural forecasting results by an ensemble system. We will integrate the results of partial predictions made by three solutions, out of which one relies on a multilayer perceptron and two others on self-organizing networks of the competitive type. As the expert system we will apply different integration methods: simple averaging, SVD based weighted averaging, principal component analysis and blind source separation. The results of numerical experiments, concerning forecasting the hourly load for the next 24 hours of the Polish power system, will be presented and discussed. We will compare the performance of different ensemble methods on the basis of the mean absolute percentage error, mean squared error and maximum percentage error. They show a significant improvement of the proposed ensemble method in comparison to the individual results of prediction. The comparison of our work with the results of other papers for the same data proves the superiority of our approach.