摘要
对于油价波动,乃至于各类市场价格(如期货、外汇、股市)的波动预测,经济学界已经有几十年的研究。然而受限于时代因素、技术因素、从业者的知识背景等因素,经济学界对于价格波动的研究主要集中在技术面,建立的模型多为基于各类结构化数据的公式型预测模型,使用机器学习手段的属于少数,而使用非结构化数据和近年来领先的深度学习技术相结合的更为稀有。从这个角度而言,本研究具有跨学科、跨领域的特点,对于此类研究有贡献意义。
For the fluctuation of oil prices, and even the volatility forecast of various market prices (such as futures, foreign exchange, stock market), the economics industry has been studying for decades. However, due to factors such as the times, technical factors, and the knowledge background of practitioners, the economics research on price fluctuations mainly focuses on the technical aspects. The established models are mostly formula-based prediction models based on various structured data, using machines. Learning methods are a minority, and the use of unstructured data combined with leading deep learning techniques in recent years is even more rare. From this perspective, this study has interdisciplinary and cross-disciplinary characteristics and contributes to such research.
作者
卢文君
LU Wenjun(Research Institute of CNPC Beijing Richfit Information Technology Co.,Ltd.,Beijing,102206,China)
出处
《数码设计》
2018年第6期24-26,共3页
Peak Data Science
关键词
原油价格预测
神经网络
非结构化数据
情感分析
机器学习
crude oil price forecast
neural network
unstructured data
sentiment analysis
machine learning