关键词: 情感分析, 小波神经网络, SA-WNN模型, WTI现货价格预测

Abstract: Crude oil price is influenced not only by the traditional demand-supply factor, but also readily by non-conventional factors such as wars, financial crisis, natural disasters and political events. This paper, aiming at forecasting the crude oil price and improving oil price forecasting theory, uses sentiment analysis (SA) to process the non-conventional text data and to estimate the market trend, which is input into wavelet neural network (WNN) to establish a SA-WNN forecasting model. Compared with the traditional BP neural network model and ICA-WNN model, SA-WNN model can forecast the general trend of oil price precisely, making it an excellent forecasting model.

Key words: sentiment analysis, WNN, SA-WNN model, WTI stock price forecast