臺灣博碩士論文加值系統

服裝業最常見的問題是產品必須事先製造，並轉移到販售店鋪進行銷售動作，而內衣業並非將所有的產品量一次性生產，而是經歷一段時間的試售期後，交由公司高層以主觀方式判定是否繼續生產該產品，而錯誤的翻單決策將導致商品的庫存居高不下，令公司整體的營運利益造成傷害。

本研究從緒論描述研究目的與動機，並藉由文獻探討決策樹、隨機森林等模型來建立客觀的分類模型，幫助案例公司在一個月的試售期後能藉由模型來判斷新產品是否熱賣能有所依據，而經過案例分析，驗證了模型的可行性，最終交由案例公司所使用。

The most common problem in the clothing industry is that the products must be manufactured in advance and transferred to the sales shop for sales. The underwear industry does not produce all the products at one time, but after a period of trial sales, it is handed over to the company. Subjectively determine whether to continue to produce the product, and the wrong decision to turn the order will lead to high inventory of goods, causing damage to the company's overall operating interests.

This study describes the purpose and motivation of the research from the introduction, and explores the decision tree and random forest model to establish an objective classification model to help the case company to forcast whether the new product is hot after the one-month trial sale period. It can be based on the case, and after the case study, the feasibility of the model is verified and finally used by the case company.

摘要 i
Abstract ii
致謝 iii
目錄 v
圖目錄 vi
表目錄 vii
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機與目的 2
1.3 研究方法 2
1.4 預期成果 3
1.5 研究限制 3
1.6 論文架構 4
第二章 文獻回顧 5
2.1 決策樹 5
2.2 隨機森林 8
2.2.1 樣本集合方法 8
2.2.2 隨機森林流程 8
2.2.3 OOB(Out Of Bag)估計 9
2.2.4 方法比較 10
2.3 文獻整理 11
第三章 模型建立與分析 12
第四章 個案資料分析與驗證 22
4.1 個案公司介紹 22
4.2 參數調整 23
4.3 個案資料分析 25
4.3.1 較佳性 25
4.3.2 案例執行 27
第五章 結論與未來方向 28
附件一 30
附件二 46

1. 鄒明城、孫志鴻(2004)，資料探勘技術在集集大地震引致山崩之研究，國立臺灣大學理學院地理學系地理學報，第36期，第117-130頁。
2. 陳雪忠(2013)，基於隨機森林的印度洋長鰭金槍魚漁場預報，海洋學報，第35卷，第1期，第158-164頁。DOI：10.3969/j.issn.0253-4193.2013.01.018.
3. 崔東文、金波(2014)，基於隨機森林回歸演算法的水生態文明綜合評價，水利水電科技進展, 34(5):56-60。DOI:10.3880/j.issn.1006-7647.2014.05.011.
4. Fact Book | FAST RETAILING CO., LTD(2018), Retrieved from https://www.fastretailing.com/eng/ir/library/factbook.html( Jun. 23, 2019)
5. C. E. Shannon(1948). A Mathematical Theory of Communication. Bell System Technical Journal. 27(3), Page 379-423. doi:10.1002/j.1538-7305.1948.tb01338.
6. B. Efron(1979). Bootstrap Methods: Another Look at the Jackknife. The Annals of Statistics. Vol 7(1), Page 1-26.
7. J. Ross Quinlan(1986). Induction of decision trees. Machine learning, 1(1), Page 81-106.
8. Pedro Domingos(1999). The Role of Occam’s Razor in Knowledge Discovery. Data Mining and Knowledge Discovery, Vol 3 (4), Page 409-425.
9. Rodriguez-Galiano, V. F.; Ghimire, B.; Rogan, J.; Chica-Olmo, M.; Rigol-Sanchez, J. P, (2012). ISPRS Journal of Photogrammetry and Remote Sensing, Volume 67, p. 93-104. DOI: 10.1016/j.isprsjprs.2011.11.002
10. Gray KR(2013). Random forest-based similarity measures for multi-modal classification of Alzheimer's disease. NeuroImage, Vol 65, Pages 167-175. DOI: 10.1016/j.neuroimage.2012.09.065.
11. Saradel Río(2014). On the use of MapReduce for imbalanced big data using Random Forest. Information Sciences, Vol 285, Pages 112-137. doi: 10.1016/j.ins.2014.03.043.
12. WU Xiaoyu,HE Jinghan,ZHANG Pei, et al.Power System Short-term Load Forecasting Based on Improved Random Forest with Grey Relation Projection(2015).Automation of Electric Power Systems ,39(12):50-55. DOI: 10.7500/AEPS20140916005.
13. Ziwei Liu( 2016), Deep Fashion: Powering Robust Clothes Recognition and Retrieval with Rich Annotations. IEEE, DOI: 10.1109/CVPR.2016.124
14. MarianaBelgiu(2016). Random forest in remote sensing: A review of applications and future directions. ISPRS Journal of Photogrammetry and Remote Sensing, Vol 114, Pages 24-31. DOI: 10.1016/j.isprsjprs.2016.01.011.
15. Leo Breiman(2001). Random Forests. Machine Learning. Vol42(1), pp5-32. DOIhttps://doi.org/10.1023/A:101093340