本論文旨在使用多種能源發電量作為預測參數，對燃料燃燒碳排放進行預測。首先，我們採用了長短期記憶(LSTM)方法構建預測模型，發現無法穩定且一致地預測結果，推測是由於深度學習模型所需的碳排年數據量不足，以致於難以捕捉其中特徵。我們另嘗試了多變數線性擬合方法，發現在模型的驗證集中獲得了僅有1.6%的平均絕對百分比誤差，顯示了該模型的精確有效性。據此，我們預測為了達成政府訂定的西元2030年碳排放減量目標，風能發電量需要從目前的自然成長率3.5%大幅提升至20%。此外，進一步探討燃料燃燒碳排放在產業住商部門、運輸部門和電力部門之間的分布情形，推測出電力部門的碳排放量將因為承接來自另兩個部門的碳排放而需持續提升，預計在西元2027年才會出現轉折。本研究除了提供預測模型為碳排放管理和能源發展方針的制定提供了參考之外，也為未來能源轉型和碳排減量目標的實現提供了實際可行的路徑。

This thesis aims to predict carbon emissions from fuel combustion using various energy generation data as parameters. We tried Long Short-Term Memory (LSTM) but found inconsistent results, likely due to insufficient carbon emission data for the deep learning model. Instead, a multivariate linear fitting method achieved a mean average absolute percentage error of only 1.6%, showing its accuracy. To reach the government's 2030 carbon reduction target, wind energy generation needs to increase from 3.5% to 20%. We also analyzed carbon emissions among industries, transportation, and power sectors, predicting a turning point around 2027. This study offers valuable insights for carbon emission management, energy policies, and a practical path for future energy transformation and carbon reduction goals.

(1) 國家發展委員會. 淨零轉型之階段目標及行動. 2022.
(2) 國家發展委員會、行政院環境保護署、經濟部、科技部、交通部、內政部、行政院農業委員會、金融監督管理委員會. 臺灣2050淨零排放路徑及策略總說明. 2022.
(3) Abdullah, L.; Pauzi, H. M. Methods in forecasting carbon dioxide emissions: a decade review. Jurnal Teknologi 2015, 75 (1), 67-82.
(4) Hong, L.-S. A case study on how to increase energy efficiency of natural gas in Taiwan. Science and Technology Division, T.-J., Ed.; 2016.
(5) BP, D. British Petroleum Statistical Review of World Energy 2021. 2021.
(6) 經濟部能源局. 經濟部能源局_發電量年資料. 2022.
(7) 林茂文. 臺灣 2050 淨零排放路徑及策略之綜析. 石油季刊 2022, 58 (2), 1-39.
(8) Hyndman, R. J.; Athanasopoulos, G. Forecasting: principles and practice; OTexts, 2018.
(9) Gunawan, J.; Huang, C.-Y. An extensible framework for short-term holiday load forecasting combining dynamic time warping and LSTM network. IEEE Access 2021, 9, 106885-106894.
(10) Galton, F. Regression towards mediocrity in hereditary stature. The Journal of the Anthropological Institute of Great Britain and Ireland 1886, 15, 246-263.
(11) Pearson, K. Mathematical contributions to the theory of evolution, On the law of ancestral heredity. Proceedings of the Royal Society of London 1898, 62 (379-387), 386-412.
(12) Suganthi, L.; Samuel, A. A. Energy models for demand forecasting—A review. Renewable and sustainable energy reviews 2012, 16 (2), 1223-1240.
(13) Harris, J. L.; Liu, L.-M. Dynamic structural analysis and forecasting of residential electricity consumption. International Journal of Forecasting 1993, 9 (4), 437-455.
(14) Egelioglu, F.; Mohamad, A.; Guven, H. Economic variables and electricity consumption in Northern Cyprus. Energy 2001, 26 (4), 355-362.
(15) Hosseini, S. M.; Saifoddin, A.; Shirmohammadi, R.; Aslani, A. Forecasting of CO2 emissions in Iran based on time series and regression analysis. Energy Reports 2019, 5, 619-631.
(16) Asumadu-Sarkodie, S.; Owusu, P. A. Recent evidence of the relationship between carbon dioxide emissions, energy use, GDP, and population in Ghana: a linear regression approach. Energy Sources, Part B: Economics, Planning, and Policy 2017, 12 (6), 495-503.
(17) Ardakani, M. K.; Seyedaliakbar, S. M. Impact of energy consumption and economic growth on CO2 emission using multivariate regression. Energy Strategy Reviews 2019, 26, 100428.
(18) Sherstinsky, A. Fundamentals of recurrent neural network (RNN) and long short-term memory (LSTM) network. Physica D: Nonlinear Phenomena 2020, 404, 132306.
(19) 陈建廷; 向阳. 深度神经网络训练中梯度不稳定现象研究综述. 软件学报 2018, 29 (7), 2071-2091.
(20) Hochreiter, S.; Schmidhuber, J. Long short-term memory. Neural computation 1997, 9 (8), 1735-1780.
(21) 華敏學. 結合多因子與風險平價模型於資產管理的應用. 2021.
(22) Ageng, D.; Huang, C.-Y.; Cheng, R.-G. A short-term household load forecasting framework using LSTM and data preparation. IEEE Access 2021, 9, 167911-167919.
(23) Altché, F.; de La Fortelle, A. An LSTM network for highway trajectory prediction. In 2017 IEEE 20th international conference on intelligent transportation systems (ITSC), 2017; IEEE: pp 353-359.
(24) Niu, H.; Zhang, Z.; Xiao, Y.; Luo, M.; Chen, Y. A study of carbon emission efficiency in Chinese provinces based on a three-stage SBM-undesirable model and an LSTM model. International Journal of Environmental Research and Public Health 2022, 19 (9), 5395.
(25) Faruque, M. O.; Rabby, M. A. J.; Hossain, M. A.; Islam, M. R.; Rashid, M. M. U.; Muyeen, S. A comparative analysis to forecast carbon dioxide emissions. Energy Reports 2022, 8, 8046-8060.
(26) Aamir, M.; Bhatti, M. A.; Bazai, S. U.; Marjan, S.; Mirza, A. M.; Wahid, A.; Hasnain, A.; Bhatti, U. A. Predicting the Environmental Change of Carbon Emission Patterns in South Asia: A Deep Learning Approach Using BiLSTM. Atmosphere 2022, 13 (12), 2011.
(27) Lin, C.-S.; Liou, F.-M.; Huang, C.-P. Grey forecasting model for CO2 emissions: A Taiwan study. Applied energy 2011, 88 (11), 3816-3820.
(28) Lotfalipour, M. R.; Falahi, M. A.; Bastam, M. Prediction of CO2 emissions in Iran using grey and ARIMA models. International Journal of Energy Economics and Policy 2013, 3 (3), 229-237.
(29) Tudor, C. Predicting the evolution of CO2 emissions in Bahrain with automated forecasting methods. Sustainability 2016, 8 (9), 923.
(30) Alam, T.; AlArjani, A. A comparative study of CO2 emission forecasting in the gulf countries using autoregressive integrated moving average, artificial neural network, and holt-winters exponential smoothing models. Advances in Meteorology 2021, 2021, 1-9.
(31) 鄧聚龍. 灰色系統基本方法 (Grey System Theory). 華中理工大學出版社: 1982.
(32) 經濟部能源局. 110年度我國燃料燃燒二氧化碳排放統計與分析. 2022.
(33) 行政院環境保護署. 我國國家溫室氣體排放清冊報告. 2022.
(34) 行政院主計總處. 歷年各季國內生產毛額依行業分. 2022.
(35) 交通部公路總局. 機動車輛登記數－按縣市別分. 2023.
(36) 經濟部能源局. 110年度全國電力資源供需報告. 2022.
(37) 交通部運輸研究所. 運輸部門歷年二氧化碳排放量推估. 2021.