非線性灰色伯努利模型應用在新興市場的二氧化碳排放量、能源消耗量和經濟成長之中長期預測研究(I)

Abstract

由於新興市場發展快速，其碳排放量已經取代成熟市場，成為未來全球碳排放的「主 力」，本計畫提出合適的非線性創新模式，利用近年資料 (2000-2008)，分析預測新興市 場的中長期二氧化碳排放量、能源消耗量和經濟成長三變量，並分三階段建構。第一階 段為收集三變量的最新資料，建構灰關聯模式，探討影響碳排放量的重要因子排序，再 利用多變量誤差修正模式，找出三變量之因果關係，以了解碳排放的主要來源。第二階 段為創新採用非線性灰預測模型，對三變量做中長期預測，直至2015年。第三階段是創 新採用非線性灰色伯努利微分方程理論模型及數值模擬方法，找出最佳之模型參數配 對，應用於三變量的中長期預測直至2015年，並和線性模型及灰預測模型做比較，評估 各模型的樣本外預測能力。 新興市場的此三變量，具有複雜的非線性行為特徵，會受到國際經濟指標、原油價 格等不確定因素影響，是典型灰色系統，尚未見文獻將非線性灰關聯模式、灰預測模型 及灰色伯努利理論模型應用在此系統上，本研究可提昇灰色系統資料及快速變化資料的 預測精準度，作為政府制定能源發展政策、提升經濟成長及抑制二氧化碳排放的重要參 考指標。由於非線性灰色模型需要跟隨樣本資料做參數調整，以逼近最佳化，因此將本 計畫分二年完成，第一年分析預測巴西與俄羅斯二個新興市場，找出最佳模型，並和臺 灣做比較，第二年分析預測中國、印度、韓國與伊朗四個新興市場並和臺灣做比較 (最 新數據顯示中國、俄羅斯、印度、韓國、巴西與伊朗等新興市場的碳排放量占全球前十 大排放國的55.70%)。

As rapid development in emerging markets, its carbon emissions have replaced the mature markets, and become the main markets of global carbon emissions. This project evaluate three factors, including energy consumption, real GDP and population all of which dynamic affected the CO2 emissions in emerging markets using grey relational analysis for the recent period 2000-2008. In order to understand the main source of carbon emissions, the causality relationships between energy consumption, real GDP and emissions are examined using multivariate error correction model. The nonlinear grey prediction model and grey Bernoulli model are applied to predict energy consumption, output and emissions for the period between 2009 and 2015. The forecasting abilities of grey Bernoulli model are compared with both the ARIMA and grey prediction models over the out-of-sample period between 2003 and 2008. The energy consumption, real GDP and emissions in emerging markets, have a complex nonlinear behavior characteristics, will be affected by uncertain factors, such as international economic indicators, crude oil prices, etc. They are typical gray system. Currently, the nonlinear grey relational model, grey prediction model or grey Bernoulli model has not been applied on this system. This research can improve the prediction accuracy for grey system information and rapid changes information. Adjustment of the gray model parameters is based on the information in the sample to approximate the optimal, so this project will require two years to complete. Brazil and Russia will be analyze in the first year, and compared with Taiwan. In the second year, China, India, South Korea and Iran will be discussed and compared with Taiwan. Latest data show that carbon emissions in emerging markets (China, Russia, India, South Korea, Brazil and Iran) accounted for the top ten emitters of 55.70%. This research can serve as important reference for energy conservation policy and environmental policy.