WCSE

Forecasting Method of Crude Oil Output Based on Adaboost_BP

Abstract— In order to improve the prediction accuracy of BP neural network, Adaboost algorithm and BP neural network are combined to propose the Adaboost_BP prediction model. The prediction model firstly preprocesses the training data set, which selects different parameters of BP neural network to construct several BP neural networks. Then, the test data set and its distribution weight are initialized. According to the Adaboost algorithm, the prediction error of each BP weak predictor is calculated and the distribution weight of the test data set is adjusted according to the error, and the weight of the predictor is obtained. Finally, all weak predictors are combined to generate a strong predictor. Based on the original data of oilfield block from 1994 to 2016, the results show that the prediction of Adaboost_BP prediction model is more accurate than the single BP neural network model, and achieves good predictive result, which validates the effectiveness of the proposed model.

Index Terms— crude oil output, adaboost algorithm, BP neural network, strong predictor

Hongtao Hu, Rong Hui
School of Computer Science Xi’an Shiyou University, CHINA
Xin Guan
Research Institute of Petroleum Exploration & Development-LangFang, CHINA

ISBN: 978-981-11-3671-9 DOI: 10.18178/wcse.2017.06.17Xsrc="http://www.wcse.org/uploadfile/2019/0823/20190823055609629.png" style="width: 120px; height: 68px;" />[Download]

Cite: Hongtao Hu, Rong Hui, Xin Guan, "Forecasting Method of Crude Oil Output Based on Adaboost_BP," Proceedings of 2017 the 7th International Workshop on Computer Science and Engineering, pp. 1192-1196, Beijing, 25-27 June, 2017.