近年來隨著電子商務蓬勃發展，電商平台日益注重提升服務品質，然而對於偏鄉地區民眾與物流服務業者而言，受限於地廣人稀且聚落分散，需求量較低而提供快速到貨的服務成本偏高，導致服務品質難以維持且成本居高不下。有鑑於此，本研究透過分析歷史電子商務訂單資料，探討需求重現性的樣態，進而建構需求預測模型，以協助物流業者判斷合適的配送時機，期能提升運輸效率性並節省物流輸配送成本，逐漸改善偏鄉物流服務的經營績效與健全經營環境，避免落入惡性循環的窘境。
本研究以個案物流公司提供之顧客訂單資料進行分析預測，針對花蓮十三鄉鎮，建構一系列需求預測模型: (1) 基礎需求預測模型，協助業者預測目標鄉鎮下一期是否有訂單的產生，作為併單配送的決策支援系統; (2) 商品數量預測模型，協助業者預測可能出現之訂單與商品數量，作為配送裝載評估的參考; (3) 預測可能的商品品項分析，將關聯性高的商品一併配送。機器學習需求預測模型，以六個機器學習方法作為基礎，搭配三種深度學習方法，評估目標鄉鎮適合哪一種機器與深度學習組合的預測準確率最高以及具有可用性。研究結果顯示，以隨機森林(Random Forest)為基礎的機器學習方式，搭配以長短期記憶(LSTM)為基礎的深度學習方法，預測準確率至少87%，能有效協助業者提供更有效率的物流輸配送服務。

In recent years, with the booming development of e-commerce, e-commerce platforms have increasingly focused on improving service quality. However, for people in rural areas and logistics service providers, due to the scarcity of people and scattered settlements, demand is low and the cost of providing fast delivery services is high, making it difficult to maintain service quality and high costs. In view of this, this study analyzes historical e-commerce order data to explore the demand reproducibility patterns and constructs a demand forecasting model to help logistics companies judge the appropriate timing of delivery in order to improve transportation efficiency and save logistics transportation and distribution costs, gradually improve the operational performance and sound business environment of rural logistics services, and avoid falling into a vicious cycle.
This study analyzes and predicts the customer order data provided by the case logistics company and constructs a series of demand forecasting models for the 13 townships in Hualien: (1) a basic demand forecasting model to help the operator predict whether there will be orders in the next period in the target townships as a decision support system for combined delivery; (2) a commodity quantity forecasting model to help the operator predict the possible orders and commodity quantities as a reference for delivery (3) Prediction of possible product item analysis, so that products with high correlation can be delivered together. The machine learning demand prediction model is based on six machine learning methods and three deep learning methods to evaluate which combination of machine and deep learning has the highest prediction accuracy and usability for the target townships. The results show that the machine learning approach based on Random Forest and the deep learning approach based on Long Short-Term Memory (LSTM) have a prediction accuracy of at least 87%, which can effectively help the industry to provide more efficient logistics delivery services.

第一章 緒論 1
1.1 研究背景與動機 1
1.2 研究目的 4
1.3 研究範圍 4
1.4 研究限制 4
1.5 研究流程 5
第二章 文獻回顧 7
2.1 電子商務 7
2.2 電子商務的不同應用 11
2.3 預測方法 12
2.4 小結 17
第三章 研究方法 19
3.1 研究課題與基本假設 19
3.2 研究方法 22
3.3 模型建構流程與求解工具 27
第四章 模型建構 31
4.1 商業分析與理解 31
4.2 資料分析與理解 32
4.3 資料預處理 39
4.4 模型建構 49
4.5 評估模型成效 56
第五章 案例分析 61
5.1 分析結果 61
5.2 超參數組合 87
5.3 評估指標 93
5.4 小結 109
第六章 結論與建議 111
6.1 結論 111
6.2 建議 112
參考文獻 113
附錄 117

Bandara, K., Bergmeir, C., & Smyl, S. (2020). Forecasting across time series databases using recurrent neural networks on groups of similar series: A clustering approach. Expert systems with applications, 140, 112896.
Bandara, K., Shi, P., Bergmeir, C., Hewamalage, H., Tran, Q., & Seaman, B. (2019). Sales demand forecast in e-commerce using a long short-term memory neural network methodology. Paper presented at the International conference on neural information processing.
Bengio, Y. (2009). Learning deep architectures for AI. Foundations and trends® in Machine Learning, 2(1), 1-127.
Bergstra, J., & Bengio, Y. (2012). Random search for hyper-parameter optimization. Journal of machine learning research, 13(2).
Bloch, M., Pigneur, Y., & Segev, A. (1996). On the road of electronic commerce-a Business value framework, Gaining competitive advantage and some research issues. Paris: Université de Lausanne, Ecole des Hautes Etudes Commerciales/Institut L’Informatique et Organisation.
Borovykh, A., Bohte, S., & Oosterlee, C. W. (2017). Conditional time series forecasting with convolutional neural networks. arXiv preprint arXiv:1703.04691.
Bosnić, Z., & Kononenko, I. (2009). An overview of advances in reliability estimation of individual predictions in machine learning. Intelligent Data Analysis, 13(2), 385-401.
Breiman, L. (2001). Random forests. Machine learning, 45(1), 5-32.
Chen, C., Liu, Z., Zhou, J., Li, X., Qi, Y., Jiao, Y., & Zhong, X. (2019). How much can a retailer sell? Sales forecasting on tmall. Paper presented at the Pacific-Asia Conference on Knowledge Discovery and Data Mining.
Chiu, D.-Y., & Lin, C.-C. (2008). Exploring internal mechanism of warrant in financial market with a hybrid approach. Expert systems with applications, 35(3), 1237-1245.
Chniti, G., Bakir, H., & Zaher, H. (2017). E-commerce time series forecasting using LSTM neural network and support vector regression. Paper presented at the Proceedings of the international conference on big data and Internet of Thing.
El-Bakry, H. M., & Zhao, Q. (2006). Fast normalized neural processors for pattern detection based on cross correlation implemented in the frequency domain. Journal of Research and Practice in Information Technology, 38(2), 151-170.
Gardner, M. W., & Dorling, S. (1998). Artificial neural networks (the multilayer perceptron)—a review of applications in the atmospheric sciences. Atmospheric environment, 32(14-15), 2627-2636.
Habla, C., Driebel, R., Monch, L., Ponsignon, T., & Ehm, H. (2007). A short-term forecast method for demand quantities in semiconductor manufacturing. Paper presented at the 2007 IEEE International Conference on Automation Science and Engineering.
Han, J., Pei, J., & Kamber, M. (2011). Data mining: concepts and techniques: Elsevier.
Hensher, D. A., & King, J. (2001). Parking demand and responsiveness to supply, pricing and location in the Sydney central business district. Transportation Research Part A: Policy and Practice, 35(3), 177-196.
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780.
Jadhav, V., CHINNAPPA, R. B., & Gaddi, G. (2017). Application of ARIMA model for forecasting agricultural prices.
Kleinbaum, D. G., Dietz, K., Gail, M., Klein, M., & Klein, M. (2002). Logistic regression: Springer.
Koutsoukas, A., Monaghan, K. J., Li, X., & Huan, J. (2017). Deep-learning: investigating deep neural networks hyper-parameters and comparison of performance to shallow methods for modeling bioactivity data. Journal of cheminformatics, 9(1), 1-13.
KumarMahto, A., Biswas, R., & Alam, M. A. (2019). Short term forecasting of agriculture commodity price by using ARIMA: based on Indian market. Paper presented at the International Conference on Advances in Computing and Data Sciences.
Law, R., & Au, N. (1999). A neural network model to forecast Japanese demand for travel to Hong Kong. Tourism Management, 20(1), 89-97.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. nature, 521(7553), 436-444.
Li, M., Ji, S., & Liu, G. (2018). Forecasting of Chinese E-commerce sales: an empirical comparison of ARIMA, nonlinear autoregressive neural network, and a combined ARIMA-NARNN model. Mathematical Problems in Engineering, 2018.
Logar, A. M. (1992). Recurrent neural networks and time series prediction.
Loureiro, A. L., Miguéis, V. L., & da Silva, L. F. (2018). Exploring the use of deep neural networks for sales forecasting in fashion retail. Decision Support Systems, 114, 81-93.
Omar, H., Hoang, V. H., & Liu, D.-R. (2016). A hybrid neural network model for sales forecasting based on ARIMA and search popularity of article titles. Computational Intelligence and Neuroscience, 2016.
Ozturk, S., & Ozturk, F. (2018). Forecasting energy consumption of Turkey by Arima model. Journal of Asian Scientific Research, 8(2), 52-60.
Pai, P.-F., & Lin, C.-S. (2005). A hybrid ARIMA and support vector machines model in stock price forecasting. Omega, 33(6), 497-505.
Qi, Y., Li, C., Deng, H., Cai, M., Qi, Y., & Deng, Y. (2019). A deep neural framework for sales forecasting in e-commerce. Paper presented at the Proceedings of the 28th ACM International Conference on Information and Knowledge Management.
Roundy, R. (2001). Report on practices related to demand forecasting for semiconductor products. Retrieved from
Salinas, D., Flunkert, V., Gasthaus, J., & Januschowski, T. (2020). DeepAR: Probabilistic forecasting with autoregressive recurrent networks. International Journal of Forecasting, 36(3), 1181-1191.
Singh, A., Thakur, N., & Sharma, A. (2016). A review of supervised machine learning algorithms. Paper presented at the 2016 3rd International Conference on Computing for Sustainable Global Development (INDIACom).
Tan, P.-N., Steinbach, M., & Kumar, V. (2018). Introduction to data mining Pearson Education India. In: Indian Nursing Council New Delhi, India.
Tiexin, C., Miaomiao, T., & Ze, M. (2012). The model of parking demand forecast for the urban CCD. Energy Procedia, 16, 1393-1400.
Wang, Y., Wang, C., Shi, C., & Xiao, B. (2018). Short-term cloud coverage prediction using the ARIMA time series model. Remote Sensing Letters, 9(3), 274-283.
Witten, I. H., & Frank, E. (2002). Data mining: practical machine learning tools and techniques with Java implementations. Acm Sigmod Record, 31(1), 76-77.
Wong, K. K. (1997). The relevance of business cycles in forecasting international tourist arrivals. Tourism Management, 18(8), 581-586.
Wu, X., Kumar, V., Ross Quinlan, J., Ghosh, J., Yang, Q., Motoda, H., . . . Yu, P. S. (2008). Top 10 algorithms in data mining. Knowledge and information systems, 14(1), 1-37.
Yuan, B., & Liu, W. (2012). Measure oriented training: a targeted approach to imbalanced classification problems. Frontiers of Computer Science, 6(5), 489-497.
Zhao, K., & Wang, C. (2017). Sales forecast in E-commerce using convolutional neural network. arXiv preprint arXiv:1708.07946.
林秀芳等. (2013). 電子商務的定義. Retrieved from http://ecommercetaiwan.blogspot.com
林庭暉. (2021). 以電商需求資料分析為基礎之偏鄉物流管理策略.
陳芳瑜. (2020). B2C需求品項數據分析_以個案物流公司為例.
黃寅翔. (2020). B2C輸配送數據分析_以個案物流公司為例.
經濟部商業司. (2012). 電子商務C2C拍賣平台交易安全規範.