基于机器学习的芳烃产品价格估算  

Aromatics Price Estimating Based on Machine Learninge-owned enterprises,and made great efforts to be a leaderin driving i andconsolidating industrial foundation.Under the background of profound changes in international and

作　　者：张芷筠 张玉琳 刘洋 Zjamg Zhiyun;Zhang Yulin;Liu Yang(SINOPEC Engineering Incorporation,Beijing 100101,China)

机构地区：[1]中国石化工程建设有限公司,北京100101

出　　处：《当代石油石化》2023年第3期44-50,共7页Petroleum & Petrochemical Today

摘　　要：运用机器学习技术与技术经济学理论知识,搭建多种机器学习模型,运用线性回归、K近邻(KNN)、支持向量机(SVM)、决策树、随机森林、神经网络6种方法,结合近10年来的芳烃–聚酯产业链市场数据,进行部分芳烃产品的价格估算。对比不同模型的估算精确度,探讨实际工作需求与可接受的估算效果。Using machine learning technology and the related theoretical knowledge in technical economics,several machine learning models were built with six methods including linear regression,K neighbor,support vector machines,decision tree,random forests and neural network.These models were used to estimate the prices of some aromatics products by using the relevant market data in the last 10 years.This paper compared the estimation accuracy of different models and discussed the actual working requirements and acceptable estimation results with technical and economic analysts.

关 键 词：机器学习 价格估算 技术经济 芳烃–聚酯产业链 

分 类 号：F767[经济管理—产业经济] F426.7[自动化与计算机技术—控制理论与控制工程] TP181[自动化与计算机技术—控制科学与工程]