臺灣博碩士論文加值系統

快時尚的服飾產業在現今的社會，已經如影隨形地充斥在生活當中，且改變人們購買服裝的方式，快時尚產業的出現，也讓過往的服飾業或紡織業的產量及市場價值迅速提升。對於產品生命週期短、產業競爭強烈和需求呈現高度隨機的快時尚服飾產業來說，產品需求的預測是很重要的。快時尚的服裝零售商總是需要提供顧客所需要的產品，且在正確的時間點提供產品，並同時保持必要的庫存，因此快時尚的製造商和零售商準確地預測需求會對於生產計劃、庫存管理、銷售策略有著極大的幫助。若能提供準確的預測方法，能夠幫助公司營運，有效地提升銷售業績，並降低庫存貨量及營運成本。
本研究使用小波轉換對快時尚的實際銷售數據進行數據的處理，希望能在高波動及不確定因素的數據當中，找到原始數據潛在的趨勢或週期訊號，並結合機器學習中的長短期記憶（Long Short-Term Memory，LSTM）和極限學習機（Extreme Learning Machine，ELM）方法來建立預測模型，再使用啟發式演算法中的基因演算法（Genetic Algorithms，GA）及粒子群最佳化演算法（Particle Swarm Optimization，PSO）找到長短期記憶模型的最佳超參數，並以最佳超參數建立預測模型。最後將四種不同的預測方法進行比較，發現PSO-LSTM在兩個不同案例上皆有較好的預測效果，而利用啟發式演算法搜尋最佳的長短期網路超參數，也會得到較好的預測效果。

The fast-fashion industry has become ubiquitous in the society nowadays, and has changed the way people purchase clothes. The emergence of the fast fashion industry has also led to a rapid increase in the output and market value of the fashion and textile industries. It is important for the fast fashion apparel industry, which has a short product life cycle, strong industry competition and highly unpredictable demand, to forecast product demand. Fast fashion retailers always need to provide the right products at the right time while maintaining necessary inventory. Therefore, accurate demand forecasting for fast fashion manufacturers and retailers greatly aids in production planning, inventory management, and sales strategies. An accurate forecasting methodology will help the operation of the company to effectively improve sales performance, reduce inventory levels and operating costs.
This study uses wavelet transform to process actual sales data from the fast fashion industry in order to identify potential trends or cyclical signals in the raw data with high volatility and uncertainty. The forecasting models are built by Long Short-Term Memory (LSTM) and Extreme Learning Machine (ELM) methods in machine learning. H The optimal hyperparameters of the LSTM are found using Genetic Algorithms (GA) and Particle Swarm Optimization (PSO) in the heuristic algorithm, and the prediction model is established with the optimal hyperparameters. Finally, the four different prediction methods are compared. It is found that PSO-LSTM has better prediction results for two different cases, and the optimal LSTM hyperparameters are searched by the heuristic algorithm, which also gives better prediction results.

摘 要 I
ABSTRACT II
致謝詞 III
目錄 V
表目錄 VII
圖目錄 IX
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 研究架構 5
第二章 文獻探討 6
2.1 快時尚 6
2.2 小波轉換 20
2.3 長短期記憶 23
2.4 極限學習機 27
2.5 啟發式演算法 31
2.6 小結 35
第三章 研究方法 36
3.1 資料蒐集 37
3.2 小波轉換 38
3.3 啟發式演算法最佳化長短期記憶 40
3.4 長短期記憶 49
3.5 極限學習機 50
3.6 結果比較 52
第四章 結果與討論 53
4.1 案例1—A產品 54
4.2 案例2—B產品 68
第五章 結論與建議 83
5.1 結論 83
5.2 未來建議 84
參考文獻 85

[1]Aftab, M. A., Yuanjian, Q., Kabir, N., and Barua, Z. (2018). Super Responsive Supply Chain: The Case of Spanish Fast Fashion Retailer Inditex-Zara. International Journal of Business and Management,13(5),212–227.
[2]Aksoy, A., Ozturk, N., and Sucky, E. (2012). A Decision Support System for Demand Forecasting in The Clothing Industry. International Journal of Clothing Science and Technology, 24(4), 221–236.
[3]Au, K. F., Choi, T. M., and Yu, Y. (2008). Fashion Retail Forecasting by Evolutionary Neural Networks. International Journal of Production Economics, 114(2), 615–630.
[4]Aufa, B. Z., Suyanto, S., and Arifianto, A. (2020, August). Hyperparameter setting of LSTM-based language model using grey wolf optimizer. In the Proceeding of the 2020 International Conference on Data Science and Its Applications (ICoDSA). Bandung, Indonesia.
[5]Cachon, G. P., and Swinney, R. (2011). The value of fast fashion: Quick response, enhanced design, and strategic consumer behavior. Management Science, 57(4), 778–795.
[6]Chang, Z., Zhang, Y., and Chen, W. (2019). Electricity Price Prediction Based on Hybrid Model of Adam Optimized LSTM Neural Network and Wavelet Transform. Energy, 187(15), 115804.
[7]Chaudhuri, K. D., and Alkan, B. (2022). A Hybrid Extreme Learning Machine Model with Harris Hawks Optimisation Algorithm: An Optimized Model for Product Demand Forecasting Applications. Applied Intelligence, 52, 11489–11505.
[8]Chen, I. F., and Lu, C. J. (2021). Demand forecasting for multichannel fashion retailers by integrating clustering and machine learning algorithms. Processes, 9(9), 1578.
[9]Chen, X., Wang, D., Gao, Y., and Tian, B. (2022). Analysis of Marketing Forecasting Model Based on Genetic Neural Networks: Taking Clothing Marketing as an Example. Wireless Communications and Mobile Computing, 2022, 2387016.
[10]Choi, T. M., Hui, C. L., Liu, N., Ng, S. F., and Yu, Y. (2014). Fast Fashion Sales Forecasting with Limited Data and Time. Decision Support Systems, 59, 84–92.
[11]Choi, T. M., Hui, C. L., Ng, C. F., and Yu, Y. (2012). Color Trend Forecasting of Fashionable Products with Very Few Historical Data. IEEE Transactions on Systems, Man and Cybernetics Part C, 42(6), 84–92.
[12]Choi, T. M., Yu, Y., and Au, K. F. (2011). A Hybrid SARIMA Wavelet Transform Method for Sales Forecasting. Decision Support Systems, 51(1), 130–140.
[13]Eid, M. M., El-Kenawy, E. S. M., Khodadadi, N., Mirjalili, S., Khodadadi, E., Abotaleb, M., Alharbi, A. H., Abdelhamid, A. A., Ibrahim, A., Amer, G. M., Kadi, A., and Khafaga, D. S. (2022). Meta-heuristic optimization of LSTM-based deep network for boosting the prediction of monkeypox cases. Mathematics, 10(20), 3845.
[14]Ensafi, Y., Amin, S. H., Zhang, G., and Shah, B. (2022). Time-Series Forecasting of Seasonal items Sales Using Machine Learning – a Comparative Analysis. International Journal of Information Management Data Insights, 2(1), 100058.
[15]Falatouri, T., Darbanian, F., Patrick, B., and Udokwu, C. (2022). Predictive Analytics for Demand Forecasting – A Comparison of SARIMA and LSTM in Retail SCM. Procedia Computer Science, 200, 993–1003.
[16]FLewis, C. D. (1982). Industrial and Business Forecasting Method. London: Butterworth.
[17]Fung, Y. N., Chan, H. L., Choi, T. M., and Liu, R. (2021). Sustainable product development processes in fashion: Supply chains structures and classifications. International Journal of Production Economics, 231, 107911.
[18]Güven, I., and Şimşir, F. (2020). Demand forecasting with color parameter in retail apparel industry using artificial neural networks (ANN) and support vector machines (SVM) methods. Computers & Industrial Engineering, 147, 106678.
[19]He, Q. Q., Wu, C. Y., and Si, Y. W. (2022). LSTM with Particle Swam Optimization for Sales Forecasting. Electronic Commerce Research and Applications, 51, 101118.
[20]Hochreiter, S., and Schmidhuber, J. (1997). Long Short-Term Memory. Neural Computation, 9(8), 1735–1780.
[21]Holland, J. H. (1975). Adaptation in Natural and Artificial Systems. Applying Genetic Algorithm to Increase the Efficiency of a Data Flow-based Test Data Generation Approach. Cambridge: The MIT Press.
[22]Hong, J. K. (2021). LSTM-based Sales Forecasting Model. KSII Transactions on Internet and Information Systems (TIIS), 15(4), 1232–1245.
[23]Hu, B., and Li, J. (2022). Prophet-LSTM Combination Model Based on PSO. International Journal of Research in Engineering and Science (IJRES), 10(4), 61–70.
[24]Huang, G. B., Zhu, Q. Y., and Siew, C. K. (2006). Extreme learning machine: theory and applications. Neurocomputing, 70(1-3), 489–501.
[25]Huang, H., and Liu, Q. (2017). Intelligent retail forecasting system for new clothing products considering stock-out. Fibres & Textiles in Eastern Europe, 1(121), 10–16.
[26]Iwabuchi, K., Kato, K., Watari, D., Taniguchi, I., Catthoor, F., Shirazi, E., and Onoye, T. (2022). Flexible electricity price forecasting by switching mother wavelets based on wavelet transform and Long Short-Term Memory. Energy and AI, 10, 100192.
[27]Kennedy, J., and Eberhart, R. (1995, November). Particle swarm optimization. In the Proceeding of the ICNN'95-International Conference on Neural Networks. Perth, Australia.
[28]Kriechbaumer, T., Angus, A., Parsons, D., and Casado, M. R. (2014). An improved wavelet-ARIMA approach for forecasting metal prices. Resources Policy, 39, 32–14.
[29]Li, Y., Yang, Y., Zhu, K., and Zhang, J. (2021). Clothing sale forecasting by a composite GRU–Prophet model with an attention meDONism. IEEE Transactions on Industrial Informatics, 17(12), 8335–8344.
[30]Liu, K., Cheng, J., and Yi, J. (2022). Copper price forecasted by hybrid neural network with Bayesian Optimization and wavelet transform. Resources Policy, 75, 102520.
[31]Lv, J., Han, S., and Hu, J. (2023). Clothing Sales Forecast Considering Weather Information: An Empirical Study in Brick-and-Mortar Stores by Machine-Learning. Journal of Textile Science and Technology, 9(1), 1–19.
[32]Mahjoub, S., Labdai, S., Chrifi-Alaoui, L., Marhic, B., and Delahoche, L. (2023). Short-Term Occupancy Forecasting for a Smart Home Using Optimized Weight Updates Based on GA and PSO Algorithms for an LSTM Network. Energies, 16(4), 1641.
[33]Mallat, S. G. (1989). A Theory for Multiresolution Signal Decomposition: The Wavelet Representation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 11(7), 674.
[34]Mostard, J., Teunter, R., and Koster, R. (2011). Forecasting demand for single-period products: A case study in the apparel industry. European Journal of Operational Research, 211(1), 139–147.
[35]Ni, Y., and Fan, F. (2011). A Two-Stage Dynamic Sales Forecasting Model for the Fashion Retail, Expert Systems with Applications, 38(3), 1529–1536.
[36]Pan, S. Y., Liao, Q., and Liang, Y. T. (2022). Multivariable sales prediction for filling stations via GA improved BiLSTM. Petroleum Science, 19(5), 2483–2496.
[37]Passoupathi, K., and Krishnan, V. R. (2016). Forecasting Sales for Indian Clothing Retail Businesses. Retrived from https://www.researchgate.net/publication/336553530
[38]Peng, L., Wang, L., Xia, D., and Gao, Q. (2022). Effective Energy Consumption Forecasting Using Empirical Wavelet Transform and Long Short-Term Memory. Energy, 238, 121756.
[39]Ramos, R., Santos, N., and Rebelo, R. (2015). Performance of State Space and ARIMA Models for Consumer Retail Sales Forecasting. Robotics and Computer-Integrated Manufacturing, 34, 151–163.
[40]Rauf, H. T., Gao, J., Almadhor, A., Arif, M., and Nafis, M. T. (2021). Enhanced bat algorithm for COVID-19 short-term forecasting using optimized LSTM. Soft Computing, 25(20), 12989–12999.
[41]Shahid, F., Zameer, A., & Muneeb, M. (2021). A novel genetic LSTM model for wind power forecast. Energy, 223, 120069.
[42]Shannon, C. E., and McCarthy, J. (1956). Automata Studies. Annals of Mathematics Studies. Princeton. New Jersey: Princeton University Press.
[43]Shao, B., Li, M., Zhao, Y., and Bian, G. (2019). Nickel price forecast based on the LSTM neural network optimized by the improved PSO algorithm. Mathematical Problems in Engineering, 2019, 1934796.
[44]Sharma, A. K., Bathula, S., and Saha, K. (2022). Method for Improvement of Product Sales Forecast for Long Horizon Using Hybrid Decomposition and Machine Learning on Multi-variate Time Series Data. ICDSMLA 2020. Singapore: Springer.
[45]Singh, P. K., Gupta, Y., Jha, N., and Rajan, A. (2019). Fashion retail: Forecasting demand for new items. ArXiv, 1907.01960.
[46]Statista. Retrived from https://www.statista.com/statistics/1008241/fast-fashion-market-value-forecast-worldwide/
[47]Suhermi, N., Suhartono, Permata, R. P., and Rahayu, S. P. (2019, August). Forecasting the Search Trend of Muslim Clothing in Indonesia on Google Trends Data Using ARIMAX and Neural Network. In the Proceeding of the 5th International Conference on Soft Computing in Data Science (SCDS 2019). Iizuka, Japan.
[48]Sun, Z. L., Choi, T. M., Au, K. F., and Yu, Y. (2008). Sales Forecasting Using Extreme Learning Machine with Applications in Fashion Retailing. Decision Support Systems,46(1), 411–419.
[49]Wong, W. K., and Guo, Z. X. (2010). A Hybrid Intelligent Model for Medium-Term Sales Forecasting in Fashion Retail Supply Chains Using Extreme Learning Machine and Harmony Search Algorithm. International Journal of Production Economics, 128(2), 614–624.
[50]Wu, J., and Wang, Z. (2022). A Hybrid Model for Water Quality Prediction Based on an Artificial Neural Network, Wavelet Transform, and Long Short-Term Memory. Decision Support Tools for Water Quality Management, 14(4), 610.
[51]Xia, M., Zhang, Y. C., Weng, L. G., and Ye, X. L. (2012). Fashion Retailing Forecasting Based on Extreme Learning Machine with Adaptive Metrics of Inputs. Knowledge-Based Systems, 36, 253–259.
[52]Xie, H. H., Li, C., Ding, N., and Kantasa-ard, C. (2022, January). Walmart Sale Forecasting Model Based on LSTM and LightGBM, In the Proceeding of 2021 2nd International Conference on Education, Knowledge and Information Management (ICEKIM). Xiamen, China.
[53]Yelland, P. M., and Dong, X. (2013). Forecasting demand for fashion goods: a hierarchical Bayesian approach. Intelligent Fashion Forecasting Systems: Models and Applications. Berlin, Heidelberg: Springer Berlin Heidelberg.
[54]Yesli, E., Kaya, M., and Siradag, S. (2012, July). Fuzzy Forecast Combiner Design for Fast Fashion Demand Forecasting, In the Proceeding of 2012 International Symposium on Innovations in Intelligent Systems and Applications. Trabzon, Turkey.
[55]Yu, X., and Lin, L. (2022). Fast Fashion Demand Forecasting Models: A Comparative Study, Operations Management in the Era of Fast Fashion. New York: Springer, 49–69.
[56]Yu, Y., Choi, T. M., and Hui, C. L. (2011). An intelligent fast sales forecasting model for fashion products. Expert Systems with Applications, 38(6), 7373–7379.
[57]Yu, Y., Hui, C. L. and Choi, T. M. (2012). An Empirical Study of Intelligent Expert Systems on Forecasting of Fashion Color Trend. Expert Systems with Applications, 39(4), 4383–4389.
[58]Zhang, H., Li, S., Cheng, Z., Tian, Z., Jin, H., and Hu, R. (2021, November). Analysis of Urban Parking Supply and Demand Based on PSO-LSTM. In the Proceeding of the 2021 International Conference on Computer, Internet of Things and Control Engineering (CITCE). Guangzhou, China.
[59]Zhang, W., Yang, K., Yu, N., Cheng, T., & Liu, J. (2020, December). Daily milk yield prediction of dairy cows based on the GA-LSTM algorithm. In the Proceeding of the 2020 15th IEEE International Conference on Signal Processing (ICSP). Beijing, China.
[60]Zhou, X., Meng, J. F., Wang, G. S., and Qin, X. X. (2020). A Demand Forecasting Model Based on the Improved Bass Model for Fast Fashion Clothing. International Journal of Clothing Science and Technology, 33(1), 106–121.
[61]Zolfaghari, M., and Golabi, M. R. (2021). Modeling and Predicting the Electricity Production in Hydropower Using Conjunction of Wavelet Transform, Long Short-Term Memory and Random Forest Models. Renewable Energy, 170, 1367–1381.