本文提出一種結合詞袋模組(Bag-of-Words model)與既有學習模組之微電網負載預測模型，其中包含脊迴歸(Ridge Regression)、核脊迴歸(Kernel Ridge Regression)及類神經網路(Neural Network)，先分析歷史負載數據歸納出其中規律性做初步的分類，再透過詞袋模組將其二次細分並重組，最後經由學習模組進行學習與預測。

在現今電力事業自由化與民營化的趨勢下，精準又有效率的負載預測模型將是未來能源管理上不可或缺的技術，微型電網的負載預測模型中，可依據不同管理需求建構對應預測週期的負載預測模型，如：超短期、短期及中期預測，超短期負載預測係指數分鐘至1日內，主要應用於安全監視及緊急狀況處理；短期負載預測為1日至1週內，一般使用在電力調度與機組經濟組合的場合；中期負載預測則為1月至1年內，大多於燃料採購計畫及機組檢修排程中使用。

基於上述應用場合，本文將預測模型分為15分鐘、1小時、1日、1週及1月之負載預測，分析詞袋模組應用於不同時間週期的負載預測之可行性，並比較各學習模組的優劣所在，最後，使用實際時間序列負載數據進行驗證。模擬結果顯出，本文所提出的方法確實可做為微型電於網各預測週期應用場合中的參考。

In this thesis, an innovative load forecasting model is developed based on combining the method of Bag-of-Words (BoW) and machine learning techniques, including Ridge Regression, Kernel Ridge Regression, and Neural Network. First of all historical load data is sorted by regularity and BoW. Then, machine learning models will be trained with historical data for load forecasting.

In the current trend of liberalization and privatization of the power industry, accurate load forecasting is an essential technique for microgrid energy management system. There are different kinds of load forecasting models based on different intervals such as very short-term (VST) model, short-term (ST) model, and mid-term (MT) model. The interval of a VST lasts for several minutes to one day. VST mainly applied to security monitoring and emergency condition. Short-term interval span from one day to a week. Generally, ST is used for power dispatching and economic scheduling of generating units. Mid-term interval generally refers to one month to a year. MT is mostly used for fuel procurement planning and maintenance of electrical equipment.

Based on the applications mentioned above, forecasting model is divided into five parts based on the following intervals: fifteen minutes ahead, one hour ahead, one day ahead, one week ahead, and one month ahead. Feasibility analysis of BoW for the forecasting model mentioned above, and comparison of proposed machine learning models are discussed. Finally, the effectiveness of approach is demonstrated on real load data which shows the applicability for a microgrid.

[1] Kwang-Ho Kim, Hyoung-sun Youn, Yong-Cheol Kang, “Short-Term Load Forecasting for Special Days in Anomalous Load Conditions using Neural Networks and Fuzzy Inference Method,” IEEE Transactions on Power Systems, vol.15, no.2, pp.559-565, 2000.
[2] Nima Amjady, “Short-Term Hourly Load Forecasting Using Time-Series Modeling with Peak Load Estimation Capability,” IEEE Transactions on Power Systems, vol.16, no.4, pp.798-805, 2001.
[3] E. Doveh, P. Feigin, D. Greig, L. Hyams, “Experience with FNN Models for Medium Term Power Demand Predictions,” IEEE Transactions on Power Systems, vol.14, no.2, pp.538-546, 1999.
[4] Zwe-Lee Gaing, Rong-Ceng Leon, “Optimal Grey Topological Predicting Approach to Short-Term Load Forecasting in Power System,” IEEE Power Engineering Society Summer Meeting, vol.3, pp.1244-1250, 2002.
[5] S.J.Kiartzis, C.E. Zoumas, J.B. Theocharis, A.G. Bakirtzis, V. Petridis, “Short-Term Load Forecasting in an Autonomous Power System Using Artificial Neural Networks,” IEEE Transactions on Power Systems, vol.12, no.4, pp.1591-1596, 1997.
[6] Haishan Wu, Xiaoling Chang, “Power Load Forecasting with Least Squares Support Vector Machines and Chaos Theory,” The Sixth World Congress on Intelligent Control and Automation, vol.1, pp.4369-4373, 2006.
[7] 王進德、蕭大全，類神經網路與模糊控制理論入門，全華科技圖書股份有限公司，92年9月。
[8] 蘇木春、張孝德，類神經網路、模糊系統以及基因演算法則，全華科技圖書股份有限公司，2006年3月。
[9] 葉怡成，類神經網路模式應用與實作，儒林圖書公司，2006年9月。
[10] 楊志民，不確定性支持向量機—算法及應用，科學出版社，2012年1月
[11] 孫向東，蛋白質結構預測︰支持向量機的應用，2008年9月。
[12] 尼德(Neter, John)，非線性迴歸與相關分析，華藝數位股份有限公司，1998年11月。
[13] 陳順宇，迴歸分析，華泰文化事業股份有限公司，1996年5月。
[14] 施萊費爾(Schleifer, Arthur)，資料分析、迴歸與預測，弘智文化出版，2000年6月。
[15] 余桂霖，多元迴歸分析，五南出版社，2012年4月。
[16] 庫特尼爾(Kutner, Michael H.)，應用線性迴歸模型，華泰文化事業股份有限公司，2005年7月。
[17] Qia Ding, “Long-Term Load Forecast Using Decision Tree Method,” IEEE PES Power Systems Conference and Exposition, pp.1541-1543, 2006.
[18] Jian-qiang Li, Cheng-lin Niu, Ji-Zhen Liu, Jun-jie Gu, “The Application of Data Mining in Electric Short-Term Load Forecasting,” Fifth International Conference on Fuzzy Systems and Knowledge Discovery, pp.519-522, 2008.
[19] Zhang Yi, Zhang Feng, Zhu Bingquan, “A Novel Ultra-Short Term Load Forecasting Method Based on Load Trend and Fuzzy C-Means Clustering Algorithm,” International Conference on Power Systems Technology, pp.1-7, 2014.
[20] Sun Qian, Yao JianGang, Jiang WenQian, Yang Shengjie, Xu ZhenChao, “Application of Improved Fuzzy C-Means Algorithm on Bad-Data Identification and Adjustment in Short-Term Load Forecasting,” International Conference on Advanced Power System Automation and Protection, vol.1, pp.568-571, 2011.
[21] Youchan Zhu, Yujun He, “Short-term Load Forecasting Model Using Fuzzy C Means Based Radial Basis Function Network,” Sixth International Conference on Intelligent Systems Design and Applications, vol.1, pp.579-582, 2006.
[22] Felix Pasila, “Multivariate Inputs for Electrical Load Forecasting on Hybrid Neuro-Fuzzy and Fuzzy C-Means Forecaster,” IEEE International Conference on Computational Intelligence, pp.2307-2312, 2008.
[23] Vijay Katkar, Surupendu Prakash Gangopadhyay, Sagar Rathod, Aakash Shetty, “Sales Forecasting Using Data Warehouse and Naïve Bayesian Classifier,” International Conference on Pervasive Computing, pp.1-6, 2015.
[24] Mingwei Wang, Jingtao Zhou, “A Bayesian Network-Based Classifier for Machining Error Prediction,” IEEE/ASME International Conference on Advanced Intelligent Mechatronics, pp.841-844, 2014.
[25] Daniel P. Stormont, “An Online Bayesian Classifier for Object Identification,” IEEE International Workshop on Safety, Security and Rescue Robotics, pp.1-5,2007.
[26] Maneesh Pawar, “Mean Shift Face Tracking with Dynamic Target Model Update Using Bayesian Skin Classifier,” IEEE International Conference on Computational Intelligence & Computing Research, pp.1-5, 2012.
[27] Tom Botterill, Steven Mills, Richard Green, “Speeded-Up Bag-of-Words Algorithm for Robot Localisation through Scene Recognition,” 23rd International Conference on Image and Vision Computing New Zealand, pp.1-6, 2008.
[28] Massimo De Santo, Paolo Napoletano, Antonio Piertosanto, Consolatina Liguori, Vincenzo Paciello, Francesco Polese, “Mixed Graph of Terms Beyond the Bags of Words Representation of a Text,” 45th Hawaii International Conference on System Science, pp.1070-7079, 2012.
[29] Zisheng Li, Jun-ichi Imai, Masahide Kaneko, “Face and Expression Recognition Based on Bag of Words Method Considering Holistic and Local Image Features,” International Symposium on Communications and Information Technologies, pp.1-6, 2010.
[30] Katsuhiko Ishiguro, Takeshi Yamada, Shoko Araki, Tomohiro Nakatani, Hiroshi Sawada, “Probabilistic Speaker Diarization With Bag-of-Words Representations of Speaker Angle Information,” IEEE Transactions on Audio, Speech, and Language Processing, vol.20, no.2, pp.447-460, 2012.
[31] Reza Bahmanyar, Shiyong Cui, Mihai Datcu, “A Comparative Study of Bag-of-Words and Bag-of-Topics Models of EO Image Patches,” IEEE Geoscience and Remote Sensing Letters, vol.12, no.6, pp.1357-1361, 2015.