Title: Expert-based text mining with Delphi method for crude oil price prediction

Authors: Somboon Chuaykoblap; Parames Chutima; Achara Chandrachai; Natawut Nupairoj

Addresses: Technopreneurship and Innovation Management Program, Graduate School, Chulalongkorn University, 1405-1409 14th Fl. Chamchuri Sq. Bldg. Payathai Rd., Patumwan, Bangkok, 10330, Thailand ' Department of Industrial Engineering, Engineering Faculty, Chulalongkorn University, Bangkok, Thailand ' Commerce and Accountancy Faculty, Chulalongkorn University, Bangkok, Thailand ' Department of Computer Engineering, Engineering Faculty, Chulalongkorn University, Bangkok, Thailand

Abstract: As crude is one of the most important commodities, the crude price forecasting has continuously been a centre of interest. The traditional techniques are focusing on econometric models which could not cope with the short term abnormality. Text data mining from news articles could be an effective method to predict the crude oil price variation caused by irregularities, but, the main issues in text mining originate from the particularities of natural language. In this research, the expert-based Delphi text mining (EDTM) is proposed to predict the movement of crude prices when the irregularity occurs. We employ the hierarchical clustering algorithm to reveal implicit knowledge hidden in news streams. Next, the Delphi method is introduced to give weighted ratings for different corresponding events extracted from the news. Finally, a comprehensive experiment is illustrated to show the effectiveness.

Keywords: rule-based expert systems; RES; Delphi method; web-based text mining; WTM; crude oil prices; crude prices; oil price prediction; knowledge management; KM; hierarchical clustering; news processing; price forecasting; implicit knowledge.

DOI: 10.1504/IJISE.2017.083045

International Journal of Industrial and Systems Engineering, 2017 Vol.25 No.4, pp.545 - 563

Received: 20 Jan 2015
Accepted: 23 Mar 2015
Published online: 19 Mar 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article