§ 瀏覽學位論文書目資料
系統識別號 U0002-0709202310042200
DOI 10.6846/tku202300642
論文名稱(中文) 基於機器學習及深度學習對月嫂與孕婦的媒合技術
論文名稱(英文) Matching technology for maternity nurses and pregnant women based on machine learning and deep learning
第三語言論文名稱
校院名稱 淡江大學
系所名稱(中文) 資訊工程學系碩士班
系所名稱(英文) Department of Computer Science and Information Engineering
外國學位學校名稱
外國學位學院名稱
外國學位研究所名稱
學年度 111
學期 2
出版年 112
研究生(中文) 陳胤元
研究生(英文) Yin-Yuan Chen
學號 610410275
學位類別 碩士
語言別 繁體中文
第二語言別
口試日期 2023-06-12
論文頁數 54頁
口試委員 指導教授 - 陳世興(shchen@mail.tku.edu.tw)
口試委員 - 張志勇( cychang@mail.tku.edu.tw)
口試委員 - 廖文華(whliao@ntub.edu.tw)
關鍵字(中) 深度學習
推薦系統
人工智慧
知識圖譜
機器學習
冷啟動
關鍵字(英) Deep learning
Recommendation System
Artificial intelligence
Knowledge Graph
Machine Learning
Cold Start
第三語言關鍵字
學科別分類
中文摘要
推薦系統的發展趨勢逐漸普及,使得市場上各種領域的推薦技術扮演了重要的角色,藉由智慧化的推薦系統協助決策判斷,同時拓展更多元的應用。然而,要做到個性化推薦,系統需要大量的用戶購買資訊。當這些資訊不足時,就可能面臨準確率下滑,形成所謂的「冷啟動問題」。這問題在數據稀少時特別明顯,使得模型難以准確判斷使用者的喜好。因此,為了解決此問題,本論文研究目的以媒合月嫂與孕產婦為例。由於許多孕產婦聘請月嫂的經驗有限,缺乏足夠的過往紀錄,這使得推薦準確率面臨挑戰。為了提升此準確率,我們進行深入的資料處理和特徵分析,並藉由知識圖譜找尋隱性關聯,以彌補數據不足。本研究提出一套結合知識圖譜、分群、深度學習和機器學習技術的方法,即使只有有限的訓練資料也能進行推薦。最終,我們期望透過這種方法能夠有效解決推薦系統的冷啟動問題,提升其準確率。
英文摘要
The development trend of recommendation systems is becoming increasingly popular, playing a pivotal role in various fields within the market. Intelligent recommendation systems assist in decision-making and pave the way for a broader range of applications. However, for personalized recommendations, a vast amount of user purchase information is required. When this information is insufficient, it could lead to a decline in accuracy, resulting in the so-called "cold start problem." This issue becomes particularly prominent when data is sparse, making it challenging for the model to accurately discern user preferences. Thus, this study aims to address this issue by using the matching of maternity nurses with pregnant women as an example. Due to many pregnant women's limited experience in hiring maternity nurses and a lack of previous records, the recommendation accuracy faces challenges. To enhance this accuracy, we delved deep into data processing and feature analysis, utilizing knowledge graphs to discover latent relationships and compensate for the data deficiency. This study proposes a method that combines knowledge graphs, clustering, deep learning, and machine learning techniques to make recommendations even with limited training data. Ultimately, we hope that this approach can effectively resolve the cold start problem in recommendation systems, boosting their accuracy.
第三語言摘要
論文目次
目錄
目錄	VI
圖目錄	VIII
表目錄	X
第一章、簡介	1
第二章、相關研究	7
2-1推薦匹配	7
2-2滿意度預測	13
第三章、背景知識	17
3-1知識圖譜鏈路預測	17
3-2XGBOOST介紹	18
3-3DCN介紹	20
第四章、系統架構	22
4-1環境與問題描述	22
4-1-1欲解決問題	22
4-1-2目標	22
4-2系統架構	23
4-2-1前處理	23
A.	資料收集	23
B.	資料前處理	24
C.	服務經驗與舊客戶資料進行視覺化分析與觀察	28
D.	特徵工程	30
4-2-2模型建構	31
A.	過濾器建構	32
B.	模型建構與訓練	33
C.	模型評估	41
第五章、實驗分析	43
5-1環境設定	43
5-2實驗數據	43
5-3實驗結果	44
第六章、結論	52
參考文獻	53

 
圖目錄
圖1、研究目標	3
圖2、研究架構	5
圖3、系統前處理	23
圖4、非結構化資料處理流程	26
圖5、整理缺失值	27
圖6、資料轉換	27
圖7、資料標準化	28
圖8、月嫂滿意度平均	29
圖9、月嫂滿意度平均	29
圖10、月嫂滿意度平均	30
圖11、特徵萃取	31
圖12、系統架構圖	32
圖13、分群方式	34
圖14、新月嫂加入分群方式	35
圖15、知識圖譜分群建構	36
圖16、知識圖譜建構內容	37
圖17、知識圖譜建構內容	38
圖18、滿意度分類器訓練	39
圖19、DCN模型架構細節	40
圖20、加權演算法	41
圖21、資料量_分群數量的Precision	47
圖22、資料量_分群數量的Recall	47
圖23、資料量_分群數量的F1-score	48
圖24、資料量_分群數量的Precision	48
圖25、資料量_分群數量的Recall	49
圖26、資料量_分群數量的F1-score	49
圖27、分群_群內距離	50
圖28、知識圖譜_月嫂數量的F1-score	50
圖29、特徵對影響度的關係圖	51

 
表目錄
表1、相關研究比較表	15
表2、模型訓練環境	43
表3、實驗參數	44
表4、混淆矩陣舉例	45

參考文獻
[1]	T. Zhou, Y. Zeng, Y. Li, Z. Jiang, Z. Liu and T. Li, "Cold-start Recommendation Method Based on Homomorphic Encryption," 2021 International Conference on Networking and Network Applications (NaNA), Lijiang City, China, 2021, pp. 461-465, doi: 10.1109/NaNA53684.2021.00086.
[2]	B. Hawashin, S. Alzubi, A. Mughaid, F. Fotouhi and A. Abusukhon, "An Efficient Cold Start Solution for Recommender Systems Based on Machine Learning and User Interests," 2020 Seventh International Conference on Software Defined Systems (SDS), Paris, France, 2020, pp. 220-225, doi: 10.1109/SDS49854.2020.9143953.
[3]	蔡聖威 and 楊振和, "線上智能媒合投保系統," Taiwan Patent TW202326538A, applied Dec. 29, 2021, issued July 1, 2023.
[4]	川村武人, "嚮導配對系統、嚮導配對方法以及電腦可讀取記憶媒體 (Guide matching system, guide matching method and program)," Taiwan Patent TW201706939A, applied Mar. 25, 2016, issued Feb. 16, 2017.
[5]	張勝凱, "師資媒合系統 (TEACHER MATCHING SYSTEM)," Taiwan Patent TW202207150A, applied Aug. 12, 2020, issued Feb. 16, 2022.
[6]	欽憶弘, 邱文卿, 劉淑茹, and 洪立屏, "師生受教需求媒合配對系統(TEACHERS AND STUDENTS DEMAND MATCHING SYSTEM)," BEST EDUCATION SERVICE & TECH. INC., Taiwan Patent TW202326587A, applied Dec. 23, 2021, issued Jul. 1, 2023.
[7]	Y. Chen, "A music recommendation system based on collaborative filtering and SVD," 2022 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS), Dalian, China, 2022, pp. 1510-1513, doi: 10.1109/TOCS56154.2022.10016210.
[8]	S. N. Parikh, J. Shah, K. Sutaria and B. Vala, "Theoretical Evaluation of Machine Learning Approaches for Hotel Recommendation," 2023 5th International Conference on Smart Systems and Inventive Technology (ICSSIT), Tirunelveli, India, 2023, pp. 1130-1137, doi: 10.1109/ICSSIT55814.2023.10061074.
[9]	L. Feng, "Design of Tourism Intelligent Recommendation Model of Mount Tai Scenic Area Based on Knowledge Graph," 2020 International Conference on E-Commerce and Internet Technology (ECIT), Zhangjiajie, China, 2020, pp. 241-244, doi: 10.1109/ECIT50008.2020.00062.
[10]	J. Yu, J. Shi, Y. Chen, W. Liu, K. Liu and Z. Xie, "Enhancing Collaborative Filtering Recommendation by User Interest Probability," 2021 4th International Conference on Artificial Intelligence and Big Data (ICAIBD), Chengdu, China, 2021, pp. 525-529, doi: 10.1109/ICAIBD51990.2021.9459028.
[11]	Q. Guo et al., "A Survey on Knowledge Graph-Based Recommender Systems," in IEEE Transactions on Knowledge and Data Engineering, vol. 34, no. 8, pp. 3549-3568, 1 Aug. 2022, doi: 10.1109/TKDE.2020.3028705.
[12]	P. Kunekar, M. Deshpande, A. Gharpure, V. Gokhale, A. Gore and H. Yadav, "Evaluating the Predictive Ability of the LightGBM Classifier for Assessing Customer Satisfaction in the Airline Industry," 2023 International Conference for Advancement in Technology (ICONAT), Goa, India, 2023, pp. 1-6, doi: 10.1109/ICONAT57137.2023.10080120.
[13]	G. Xiao, T. -W. Kuan, F. Li, M. Ling and S. -P. Tseng, "Systematic Review of Machine Learning Methods in Customer Satisfaction Prediction," 2022 10th International Conference on Orange Technology (ICOT), Shanghai, China, 2022, pp. 1-3, doi: 10.1109/ICOT56925.2022.10008132.
[14]	H. Gou, L. Su, G. Zhang, W. Huang, Y. Rao and Y. Yang, "A XGBoost Method Based on Telecom Customer Satisfaction Enhancement Strategy," 2022 5th International Conference on Pattern Recognition and Artificial Intelligence (PRAI), Chengdu, China, 2022, pp. 209-213, doi: 10.1109/PRAI55851.2022.9904203.
[15]	ZHANG, Muhan; CHEN, Yixin. Link prediction based on graph neural networks. Advances in neural information processing systems, 2018, 31
[16]	Chen, T., & Guestrin, C. (2016, August). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining (pp. 785-794).
[17]	WANG, Ruoxi, et al. Dcn v2: Improved deep & cross network and practical lessons for web-scale learning to rank systems. In: Proceedings of the web conference 2021. 2021. p. 1785-1797.
論文全文使用權限
國家圖書館
不同意無償授權國家圖書館
校內
校內紙本論文立即公開
電子論文全文不同意授權
校內書目立即公開
校外
不同意授權予資料庫廠商
校外書目立即公開

如有問題,歡迎洽詢!
圖書館數位資訊組 (02)2621-5656 轉 2487 或 來信