基于自适应网络的模糊推理系统(ANFIS)的中国碳市场收益率预测  

Adaptive Network-Based Fuzzy Inference System (ANFIS) for Forecasting Carbon Market Yield in China

作　　者：任航 

机构地区：[1]西华大学理学院,四川 成都

出　　处：《统计学与应用》2025年第2期245-262,共18页Statistical and Application

基　　金：国家自然科学基金(项目批准号:52472368)。

摘　　要：在“双碳”背景下,中国碳排放权交易市场的收益率存在显著的动态非线性、混沌性特征,如何准确地预测碳市场收益率,促进我国绿色可持续发展越来越受到关注。本文旨在探究碳市场的交易特征和经济政策不确定性对中国碳市场收益率的影响,构建中国碳市场的自适应模糊–神经网络(ANFIS)模型预测中国碳市场收益率,进而探索碳价格的形成机制。研究发现:首先,基于中国碳市场的交易特征和经济政策不确定性所建立的自适应ANFIS模型具有较好的预测性能,尤其是具有较高的样本外预测精度,并能有效地捕获碳市场的非线性特征。其次,中国碳试点市场由于其自身试点运行机制的不同,市场的交易特征尤其是市场碎片化和经济政策不确定性对各试点的影响具有显著的异质性特征。政府在运行全国统一的碳排放权交易市场时,需要考虑市场碎片化和经济政策不确定性对碳排放权收益率的影响,确保充分发挥碳市场交易的信号功能,引导企业以最小成本实现减排目标。Under the background of “dual-carbon”, the yield of China’s carbon emissions trading market is characterized by significant dynamic nonlinearity and chaos, and how to accurately predict the yield of the carbon market to promote China’s green and sustainable development is receiving more and more attention. The purpose of this paper is to explore the impact of the trading characteristics of the carbon market and economic policy uncertainty on the yield of China’s carbon market, to construct an Adaptive Network-based Fuzzy Inference System (ANFIS) model of China’s carbon market to predict the yield of China’s carbon market, and then to explore the mechanism of the formation of carbon price. It is found that, firstly, the adaptive ANFIS model constructed based on the trading characteristics and economic policy uncertainty of the Chinese carbon market has better forecasting performance, es

关 键 词：碳市场收益率 自适应模糊神经网络 经济政策不确定性 市场碎片化 市场流动性 

分 类 号：F42[经济管理—产业经济]