基於定制服務的產品越來越多，產品項次的增加和變化影響產品生命週期管理，產品之間的替代效應也會影響實際銷售。在人力資源有限的情況下，為每個產品逐一建立銷售預測模型是一項耗時的任務，單個產品的預測總和不一定等於所有產品的預測。本論文的研究動機是針對上述商業實踐中的挑戰，提出一個可執行的解決方案。
本論文提出兩個研究目標，首先是在經典BCG矩陣的基礎上，以產品組合形式為多樣產品建立銷售預測模型，取代原來的單一產品預測模式。其次，建立一項基於集群的多樣產品銷售預測的解決方案框架。
本論文提出了一個包含研究方法和工具的解決方案框架，並通過實證案例研究來實踐研究目標。研究方法包括單一產品時間序列預測模型，BCG矩陣與產品組合分析，以及基於集群的多樣產品預測建模。
實證案例研究有三個主要結果。首先是為熱銷產品創建時間序列預測分析，主要輸出包括為公司提供推薦模型以及備選模型做為靈活決策，並採用預期差距對下個季度的生產提出建議。同時，對建模過程中採用的預測器、缺失數據處理方法以及其組合，就預測結果進行解讀。其次為熱銷產品的BCG矩陣分析，分析結果包括：LSTM 和零值填充的組合模型適合老狗產品的銷售預測、零值填充方法適用於高市佔率產品(明星與金牛產品)的缺失值處理，以及均值插補方法適用於平均市佔率產品(老狗及問題小孩產品)的缺失值處理。
第三項結果是引入基於數據驅動特徵的k-均值集群算法，對BCG矩陣進行基於科學的修正。k-均值集群算法的優點是更能敏感地反映業務特徵，有效地將包含市場策略在內的設計方案擴展到其他產品上進行部署，實現多樣產品預測的目的。此外，一個關鍵且重要的貢獻是增加了分析所需的數據使用量。本論文的主要貢獻是為多樣產品銷售預測開發了一個實用的框架，並為商業從業者提供了一個有效、穩定、可持續和可擴展的工作流程。

There are more and more products based on customized services. The increase and change of product items affect the product life cycle management, and the substitution effect between products will also affect actual sales. In the case of limited human resources, building a sales forecast model for each product one by one is a time-consuming task, and the sum of the forecasts for a single product does not necessarily equal the estimates for all products. The research motivation of this thesis is to propose an executable solution to the above challenges in business practice. This thesis proposes two research objectives. The first is to establish a sales forecasting model for the multi-item products in the form of a portfolio based on the classic BCG matrix, replacing the original single product forecast model. Secondly, to establish a solution framework for cluster-based multi-item products sales forecasting. The research methods include single-product time series forecasting models, BCG matrix and product portfolio analysis, and cluster-based multi-item products forecasting modeling.
There are three primary outcomes of the empirical case study. The first is creating a time series forecast analysis for the hot-selling products. The key results include providing the options of recommended and backup models for the company for flexible decision-making, describing the expected gaps and production recommendations for the next quarter, and providing general judgments on forecasters, missing data handling methods, and combinations thereof. The second outcome is the BCG matrix analysis of the top 10 products. The key results are that the LSTM + zero-filling model is suitable for Dogs products, the zero-filling method is ideal for high market share products (i.e., Stars and Cash-cow products), and the mean-impute is ideal for the average products (i.e., Dogs and Problem-child products). The third outcome of the empirical case study is to produce a science-based revision to the BCG matrix by introducing the k-Means Clustering based on its data-driven characteristics. The advantages of k-Means Clustering are that it can reflect business characteristics more sensitively than BCG matrix analysis, effectively expand the designed schemes to other products for deployment, and then implement the purpose of multi-item products forecasting. It also can be feasibly to develop sales forecasting schemes, including marketing strategies. Furthermore, a key and significant contribution is the increased use of data for analysis.
The primary contributions of this thesis are developing a practical framework for multi-item products sales forecasting and providing a valid, stable, sustainable, and expandable workflow for business practitioners.

Abbasimehr, H., & Khodizadeh Nahari, M. (2020). Improving demand forecasting with LSTM by taking into account the seasonality of data. Journal of Applied Research on Industrial Engineering, 7(2), 177-189.
Abbasimehr, H., Shabani, M., & Yousefi, M. (2020). An optimized model using LSTM network for demand forecasting. Computers & Industrial Engineering, 143(2020), 106435, 1-13.
Abdulla, S. H., Sagheer, A. M., & Veisi, H. (2022). Breast cancer segmentation using K-means clustering and optimized region-growing technique. Bulletin of Electrical Engineering and Informatics, 11(1), 158-167.
Ali, Ö. G., Sayın, S., van Woensel, T., & Fransoo, J. (2009). SKU demand forecasting in the presence of promotions. Expert Systems with Applications, 36(10), 12340-12348.
Altini, N., De Giosa, G., Fragasso, N., Coscia, C., Sibilano, E., Prencipe, B., Hussain, S. M., Brunetti, A., Buongiorno, D., Guerriero, A., Tatò, I. S., Brunetti, G., Triggiani, V., & Bevilacqua, V. (2021). Segmentation and identification of vertebrae in CT scans using CNN, k-means clustering and k-NN. Informatics, 8(2), 40, 1-17.
Armstrong, J. S., & Collopy, F. (1992). Error measures for generalizing about forecasting methods: Empirical comparisons. International Journal of Forecasting, 8(1), 69-80.
Arumugam, P., & Saranya, R. (2018). Outlier detection and missing value in seasonal ARIMA model using rainfall data. Materials Today: Proceedings, 5(1), 1791-1799.
Babai, M. Z., Ali, M. M., & Nikolopoulos, K. (2012). Impact of temporal aggregation on stock control performance of intermittent demand estimators: Empirical analysis. Omega, 40(6), 713-721.
Babu, C. N., & Reddy, B. E. (2014). A moving-average filter based hybrid ARIMA-ANN model for forecasting time series data. Applied Soft Computing, 23, 27-38.
Bala, P. K. (2012). Improving inventory performance with clustering based demand forecasts. Journal of Modelling in Management, 7(1), 23-37.
Barksdale, H. C., & Harris Jr, C. E. (1982). Portfolio analysis and the product life cycle. Long Range Planning, 15(6), 74-83.
Borges, C. E., Kamara-Esteban, O., Castillo-Calzadilla, T., Andonegui, C. M., & Alonso-Vicario, A. (2020). Enhancing the missing data imputation of primary substation load demand records. Sustainable Energy, Grids and Networks, 23 (2020), 100369, 1-9.
Box, G. E., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time Series Analysis: Forecasting and Control. John Wiley & Sons Inc., Hoboken, New Jersey.
Bunn, D. W., & Vassilopoulos, A. I. (1999). Comparison of seasonal estimation methods in multi-item short-term forecasting. International Journal of Forecasting, 15(4), 431-443.
Carbonneau, R., Laframboise, K., & Vahidov, R. (2008). Application of machine learning techniques for supply chain demand forecasting. European Journal of Operational Research, 184(3), 1140-1154.
Chen, C., Hu, J., Meng, Q., & Zhang, Y. (2011). Short-time traffic flow prediction with ARIMA-GARCH model. Proceedings of the Intelligent Vehicles Symposium 4, 607-612, IEEE. Baden-Baden, Germany.
Chujai, P., Kerdprasop, N., & Kerdprasop, K. (2013). Time series analysis of household electric consumption with ARIMA and ARMA models. Proceedings of the International MultiConference of Engineers and Computer Scientists 1, 295-300, IMECS. Hong Kong.
Deb, C., Zhang, F., Yang, J., Lee, S. E., & Shah, K. W. (2017). A review on time series forecasting techniques for building energy consumption. Renewable and Sustainable Energy Reviews, 74, 902-924.
Emmanuel, T., Maupong, T., Mpoeleng, D., Semong, T., Mphago, B., & Tabona, O. (2021). A survey on missing data in machine learning. Journal of Big Data, 8(1), 1-37.
Gardner, G., Harvey, A. C. & Phillips, G. D. A (1980). An algorithm for exact maximum likelihood estimation of autoregressive-moving average models by means of Kalman filtering. Applied Statistics, 29(3), 311-322.
Gómez-Carracedo, M. P., Andrade, J. M., López-Mahía, P., Muniategui, S., & Prada, D. (2014). A practical comparison of single and multiple imputation methods to handle complex missing data in air quality datasets. Chemometrics and Intelligent Laboratory Systems, 134, 23-33.
Gómez, V., Maravall, A., & Pena, D. (1999). Missing observations in ARIMA models: Skipping approach versus additive outlier approach. Journal of Econometrics, 88(2), 341-363.
Goodwin, P., & Lawton, R. (1999). On the asymmetry of the symmetric MAPE. International Journal of Forecasting, 15(4), 405-408.
Gopagoni, D. R., Lakshmi, P. V., & Chaudhary, A. (2021). Evaluating machine learning algorithms for marketing data analysis: predicting grocery store sales. Communication Software and Networks, 134, 155-163.
Hair, J. F., Black, W. C., Babin, B. J. & Anderson, R. E. (2010). Multivariate Data Analysis. 7th edition. Pearson Education Limited, Harlow, London.
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735, 1-32.
Hosseini, S. M. S., Maleki, A., & Gholamian, M. R. (2010). Cluster analysis using data mining approach to develop CRM methodology to assess the customer loyalty. Expert Systems with Applications, 37(7), 5259-5264.
Hung, C. Y., Jiang, B. C., & Wang, C. C., (2020). Evaluating machine learning classification using sorted missing percentage technique based on missing data. Applied Sciences, 10(14), 4920, 1-16.
Hung, C. Y., Wang, C. C., Lin, S. W., & Jiang, B. C. (2022). An empirical comparison of the sales forecasting performance for plastic tray manufacturing using missing data. Sustainability, 14(4), 2382, 1-21.
Hyndman, R. J., & Athanasopoulos, G. (2018). Forecasting: principles and practice, 2nd edition. OTexts, Melbourne, Australia.
Jebarani, P. E., Umadevi, N., Dang, H., & Pomplun, M. (2021). A novel hybrid k-means and GMM machine learning model for breast cancer detection. IEEE Access, 9, 146153-146162.
Johnston, F. R., Boyland, J. E., Meadows, M., & Shale, E. (1999). Some properties of a simple moving average when applied to forecasting a time series. Journal of the Operational Research Society, 50(12), 1267-1271.
Jones, R. H. (1980). Maximum likelihood fitting of ARMA models to time series with missing observations. Technometrics, 22(3), 389-395.
Junger, W. L., & De Leon, A. P. (2015). Imputation of missing data in time series for air pollutants. Atmospheric Environment, 102, 96-104.
Junninen, H., Niska, H., Tuppurainen, K., Ruuskanen, J., & Kolehmainen, M. (2004). Methods for imputation of missing values in air quality data sets. Atmospheric Environment, 38(18), 2895-2907.
Khan, A. R., Khan, S., Harouni, M., Abbasi, R., Iqbal, S., & Mehmood, Z. (2021). Brain tumor segmentation using k‐means clustering and deep learning with synthetic data augmentation for classification. Microscopy Research and Technique, 84(7), 1389-1399.
Kiefer, D., Grimm, F., Bauer, M., & van Dinther, C. (2021). Demand forecasting intermittent and lumpy time series: comparing statistical, machine learning and deep learning methods. Proceedings of the 54th Hawaii International Conference on System Sciences, 1425-1434, HICSS. Grand Wailea, Maui, Hawaii.
Kohn, R., & Ansley, C. F. (1986). Estimation, Prediction, and Interpolation for ARIMA Models with Missing Data. Journal of the American Statistical Association, 81(395), 751-761.
Kourentzes, N. (2013). Intermittent demand forecasts with neural networks. International Journal of Production Economics, 143(1), 198-206.
Lau, H. C., Ho, G. T., & Zhao, Y. (2013). A demand forecast model using a combination of surrogate data analysis and optimal neural network approach. Decision Support Systems, 54(3), 1404-1416.
Li, C., & Lim, A. (2018). A greedy aggregation–decomposition method for intermittent demand forecasting in fashion retailing. European Journal of Operational Research, 269(3), 860-869.
Liu, P., Ming, W., & Hu, B. (2021). Sales forecasting in rapid market changes using a minimum description length neural network. Neural Computing and Applications, 33(3), 937-948.
Martínez, F., Frías, M. P., Pérez-Godoy, M. D., & Rivera, A. J. (2018). Dealing with seasonality by narrowing the training set in time series forecasting with kNN. Expert Systems with Applications, 103, 38-48.
Ma, Y., Wang, N., Che, A., Huang, Y., & Xu, J. (2013). The bullwhip effect on product orders and inventory: a perspective of demand forecasting techniques. International Journal of Production Research, 51(1), 281-302.
McKnight, P. E., McKnight, K. M., Sidani, S., & Figueredo, A. J. (2007). Missing data: A gentle introduction. Guilford Press, New York, New York.
Mohajan, H. K. (2018). An Analysis on BCG Growth Sharing Matrix. Noble International Journal of Business and Management Research, 2(1), 1-6.
Moritz, S., Sardá, A., Bartz-Beielstein, T., Zaefferer, M., & Stork, J. (2015). Comparison of different methods for univariate time series imputation in R. arXiv preprint arXiv:1510.03924, 1-20.
Moshrefi, S., Abdoli, S., Kara, S., & Hauschild, M. (2020). Product portfolio analysis towards operationalising science-based targets. Procedia CIRP, 90, 377-382.
Musial, J. P., Verstraete, M. M., & Gobron, N. (2011). Comparing the effectiveness of recent algorithms to fill and smooth incomplete and noisy time series. Atmospheric Chemistry and Physics, 11(15), 7905-7923.
Musil, C. M., Warner, C. B., Yobas, P. K., & Jones, S. L. (2002). A comparison of imputation techniques for handling missing data. Western Journal of Nursing Research, 24(7), 815-829.
Nowak, M., Mierzwiak, R., Wojciechowski, H., & Delcea, C. (2020). Grey portfolio analysis method. Grey Systems: Theory and Application, 10(4), 439-454.
Posch, K., Truden, C., Hungerländer, P., & Pilz, J. (2022). A Bayesian approach for predicting food and beverage sales in staff canteens and restaurants. International Journal of Forecasting, 38(1), 321-338.
Pradana, M. G., & Ha, H. T. (2021). Maximizing strategy improvement in mall customer segmentation using k-means clustering. Journal of Applied Data Sciences, 2(1), 19-25.
Romeijnders, W., Teunter, R., & van Jaarsveld, W. (2012). A two-step method for forecasting spare parts demand using information on component repairs. European Journal of Operational Research, 220(2), 386-393.
Rubin, D. B., & Little, R. J. (2019). Statistical analysis with missing data. John Wiley & Sons Inc., Hoboken, New Jersey.
Scheffer, J. (2002), Dealing with missing data. Research Letters in the Information and Mathematical Sciences, 3(1), 153-160
Siami-Namini, S., Tavakoli, N., & Namin, A. S. (2018). A comparison of ARIMA and LSTM in forecasting time series. Proceedings of the 17th IEEE International Conference on Machine Learning and Applications, 1394-1401, IEEE. Orlando, Florida.
Sohrabpour, V., Oghazi, P., Toorajipour, R., & Nazarpour, A. (2021). Export sales forecasting using artificial intelligence. Technological Forecasting and Social Change, 163, 120480, 1-10.
Spedding, T. A., & Chan, K. K. (2000). Forecasting demand and inventory management using Bayesian time series. Integrated Manufacturing Systems, 11(5), 331-339.
Stummer, C., & Heidenberger, K. (2003). Interactive R&D portfolio analysis with project interdependencies and time profiles of multiple objectives. IEEE Transactions on Engineering Management, 50(2), 175-183.
Syakur, M. A., Khotimah, B. K., Rochman, E. M. S., & Satoto, B. D. (2018). Integration k-means clustering method and elbow method for identification of the best customer profile cluster. Proceedings of the 2nd International Conference on Vocational Education and Electrical Engineering, 336(1), 012017, 1-7, IOP Publishing. Surabaya, Indonesia.
Taylor, J. W. (2011). Multi-item sales forecasting with total and split exponential smoothing. Journal of the Operational Research Society, 62(3), 555-563.
Tayman, J., & Swanson, D. A. (1999). On the validity of MAPE as a measure of population forecast accuracy. Population Research and Policy Review, 18(4), 299-322.
Teunter, R. H., & Duncan, L. (2009). Forecasting intermittent demand: a comparative study. Journal of the Operational Research Society, 60(3), 321-329.
Tian, K., Li, J., Zeng, J., Evans, A., & Zhang, L. (2019). Segmentation of tomato leaf images based on adaptive clustering number of k-means algorithm. Computers and Electronics in Agriculture, 165, 104962. 1-7.
Tony, A., Kumar, P., Jefferson, R., & Subramanian (2021). A study of demand and sales forecasting model using machine learning algorithm. Psychology and Education Journal, 58(2), 10182-10194.
Tripathi, A. K., Saini, H., & Rathee, G. (2022). Futuristic prediction of missing value imputation methods using extended ANN. International Journal of Business Analytics, 9(3), 1-12.
Udo-Imeh, P. T., Edet, W. E., & Anani, R. B. (2012). Portfolio analysis models: a review. European Journal of Business and Management, 4(18), 101-120.
Velicer, W. F., & Colby, S. M. (2005). A comparison of missing-data procedures for ARIMA time-series analysis. Educational and Psychological Measurement, 65(4), 596-615.
Wang, C. C., Chien, C. H., & Trappey, A. J. (2021). On the application of ARIMA and LSTM to predict order demand based on short lead time and on-time delivery requirements. Processes, 9(7), 1157, 1-19
White, I. R., & Carlin, J. B. (2010). Bias and efficiency of multiple imputation compared with complete‐case analysis for missing covariate values. Statistics in Medicine, 29(28), 2920-2931.
Wirth, R., & Hipp, J. (2000). CRISP-DM: Towards a standard process model for data mining. Proceedings of the 4th International Conference on the Practical Applications of Knowledge Discovery and Data Mining 1, 29-39. New York, New York.
Wongoutong, C. (2021). Imputation methods in time series with a trend and a consecutive missing value pattern. Thailand Statistician, 19(4), 866-879.
Wu, S., Yau, W. C., Ong, T. S., & Chong, S. C. (2021). Integrated churn prediction and customer segmentation framework for telco business. IEEE Access, 9, 62118-62136.
Xiahou, X., & Harada, Y. (2022). B2C e-commerce customer churn prediction based on k-means and SVM. Journal of Theoretical and Applied Electronic Commerce Research, 17(2), 458-475.
Xie, J., Lee, T. S., & Zhao, X. (2004). Impact of forecasting error on the performance of capacitated multi-item production systems. Computers & Industrial Engineering, 46(2), 205-219.
Yuan, X., Zhang, X., & Zhang, D. (2020). Analysis of the impact of different forecasting techniques on the inventory bullwhip effect in two parallel supply chains with a competition effect. Journal of Engineering, 2020, 1-28.
Zakaria, N. A., & Noor, N. M. (2018). Imputation methods for filling missing data in urban air pollution data form Malaysia. Urbanism, 9(2), 159-166.
Zhang, H., & Peng, Q. (2022). PSO and K-means-based semantic segmentation toward agricultural products. Future Generation Computer Systems, 126, 82-87.
Zhang, Y. (2021). Sales forecasting of promotion activities based on the cross-industry standard process for data mining of e-commerce promotional information and support vector regression. Journal of Computers, 32(1), 212-225.
Zhang, Y., & Thorburn, P. J. (2022). Handling missing data in near real-time environmental monitoring: A system and a review of selected methods. Future Generation Computer Systems, 128, 63-72.