On Prediction Of EU Carbon Price Based On Improved Fruit Fly Optimization Algorithm

In order to alleviate the deteriorating environmental problems,the Kyoto Protocol formulated in Kyoto,Japan,has pushed the carbon emission rights market to a realistic stage in 1997.As a derivative market of carbon spot market,carbon futures market theoretically has the function of price prediction and risk aversion,which has a decisive impact on the rational allocation of carbon resources,the reduction of risk and the reduction of greenhouse gas emissions.EU Emission Trading System(EU ETS)is the earliest carbon emission reduction market in the world.This paper studies the international carbon futures market by taking the main carbon emission contract futures under EU ETS-EU carbon emission allowance futures and certified emission reduction futures as the research objects.The research on price trend will not only help investors make reasonable decisions,but also help to further play the role of carbon emission reduction.Moreover,it provides the reference for China to establish a unified carbon emission reduction market.This paper mainly studies the main carbon emission contract futures under EU ETS-EU carbon emission allowance futures and certified emission reduction futures.The full text is divided into five parts: The first part introduces the research significance,literature review and the main content of this paper.The second part describes the related theories of fruit fly optimization algorithm and machine learning.In the third part,the fruit fly optimization algorithm is easy to fall into the premature convergence and low precision.The traditional fruit fly optimization algorithm is improved from chaos theory,Lévy flight and adaptive step size.Then,the improved fruit fly optimization is applied to optimize the weight and threshold of BP neural network and the important parameters such as smoothing parameters and penalty factors of support vector machine.The fourth part,considering the relationship between supply factors,domestic energy,international energy,economic factors,exchange rate and climate factors and carbon price,using Lasso variable selection method for variable selection and parameter estimation.The model is used for the forecast of daily trading settlement prices for EU carbon emission allowance futures and certified emission reduction futures.The experimental results show that the carbon price model based on the improved fruit fly optimization algorithm has higher prediction accuracy and faster convergence speed than the traditional machine learning model.The main contributions of this paper are as follows:(1)Lasso variable selection method is used to select variables and estimate parameters,and the influencing factors of carbon emission trading price and the degree of influence are explored;(2)Three methods to improve the fruit fly optimization algorithm are proposed to overcome the shortcomings of the fruit fly optimization algorithm,such as premature convergence and low precision.The experimental results show that the improved fruit fly optimization algorithm has remarkable improvement in convergence accuracy.The improved fruit fly optimization algorithm is used to optimize the important parameters of BP neural network and support vector machine,and it can avoid the blindness of parameter setting.