隨著智慧城市和智能交通系統的發展，停車推薦系統在城市管理中越來越重要。然而，傳統的停車推薦方法往往忽略了使用者的個性化需求以及外部環境因素，如天氣、溫度、停車費和使用者的職業與薪資等。為了解決這些問題，本研究提出了一個基於對比學習和知識圖譜的停車推薦系統。該系統透過對客戶嵌入和停車場嵌入進行建模，並綜合考慮環境溫度等影響因素，利用知識圖譜和對比學習技術來計算用戶對各停車場的滿意度，從而精準地為用戶推薦最適合的停車場。相比傳統方法，該系統能更有效地捕捉使用者喜好與實際需求，提高推薦結果的準確性和個性化水平。本研究的目標是設計出一個更加智能化和個性化的停車推薦系統，為用戶提供更優質的停車體驗，並提升城市停車管理的效率。
隨著智慧城市和智能交通系統的發展，停車推薦系統在城市管理中越來越重要。然而，傳統的停車推薦方法往往忽略了使用者的個性化需求以及外部環境因素，如天氣、溫度、停車費和使用者的職業與薪資等。為了解決這些問題，本研究提出了一個基於對比學習和知識圖譜的停車推薦系統。該系統透過對客戶嵌入和停車場嵌入進行建模，並綜合考慮環境溫度等影響因素，利用知識圖譜和對比學習技術來計算用戶對各停車場的滿意度，從而精準地為用戶推薦最適合的停車場。相比傳統方法，該系統能更有效地捕捉使用者喜好與實際需求，提高推薦結果的準確性和個性化水平。本研究的目標是設計出一個更加智能化和個性化的停車推薦系統，為用戶提供更優質的停車體驗，並提升城市停車管理的效率。

As smart cities and intelligent transportation systems develop, parking recommendation systems are becoming increasingly important in urban management. However, traditional parking recommendation methods often overlook users' personalized needs and external factors, such as weather, temperature, parking fees, and users' occupations and salaries. To address these issues, this study proposes a parking recommendation system based on contrastive learning and knowledge graphs. The system models customer embeddings and parking lot embeddings while considering factors like environmental temperature. It utilizes knowledge graphs and contrastive learning techniques to calculate user satisfaction with different parking lots, thus providing precise recommendations for the most suitable parking options. Compared to traditional methods, this system more effectively captures user preferences and actual needs, enhancing the accuracy and personalization of the recommendations. The goal of this study is to design a more intelligent and personalized parking recommendation system, offering users a better parking experience and improving the efficiency of urban parking management.

目錄
目錄	VIII
圖目錄	VIII
表目錄	IX
第一章、簡介	1
第二章、相關研究	5
第三章、背景知識	18
3-1 EMBEDDING介紹	18
3-2 LIGHTGCN 介紹	20
3-3對比學習 介紹	22
第四章、系統架構	25
4-1環境與問題描述	25
4-1-1背景與動機	25
4-1-2目標	26
4-2整體系統架構	27
4-3 資料收集與前處理	27
4-4 嵌入生成與學習（LIGHTGCN）	29
4-5 推薦策略計算	31
4-6 出具推薦的可解釋性	35
第五章、實驗分析	39
5-1環境設定	39
5-2實驗數據	39
5-3實驗結果	40
5-4討論	43
第六章、結論	44
參考文獻	45

 
圖目錄
圖1 、整體系統架構	27
圖2 、資料前處理	28
圖3 、資料集: 建德市停車記錄原始資料	29
圖4 、核心功能架構圖	33
圖5 、用戶信息GCN嵌入示範	33
圖6 、用戶需求GCN嵌入示範	34
圖7 、停車場GCN嵌入示範	34
圖8 、上下文GCN嵌入示範	35
圖9 、用戶對比學習過程示範	35
圖10 、新用戶推薦可解釋過程	38
圖11 、新停車場推薦可解釋過程	38
圖12 、各模型的評估結果	42
圖13 、某用戶停車習慣反映在溫度天氣以及星期	43

 
表目錄

表1 、相關研究比較表	16
表2 、模型訓練環境	39
表3 、混淆矩陣舉例	41
表4 、各模型的評估結果	42

[1] S. Shukla · R. Gupta · S. Garg · S. Harit · R. Khan (*) “Real-Time 
Parking Space Detection and Management with Artifcial Intelligence 
and Deep Learning System” Department of Computer Science and 
Engineering, ABES Institute of Technology, AKTU, Ghaziabad, UP, 
India. Transforming Management with AI, Big-Data, and 
IoT,https://doi.org/10.1007/978-3-030-86749-2_7.
[2] Akram Elomiya, Jiří Křupka, Stefan Jovčić, A Smart Parking System 
Using Surveillance Cameras and Fuzzy Logic: A Case Study at 
Computer Science,https://doi.org/10.1016/j.procs.2023.10.488.
[3] S. R. Rizvi, S. Zehra and S. Olariu, "ASPIRE: An Agent-Oriented 
Smart Parking Recommendation System for Smart Cities," in IEEE 
Intelligent Transportation Systems Magazine, vol. 11, no. 4, pp. 48-
61, winter 2019, doi: 10.1109/MITS.2018.2876569.
[4] Z. Li, M. Alazab, S. Garg and M. S. Hossain, "PriParkRec: PrivacyPreserving Decentralized Parking Recommendation Service," in 
IEEE Transactions on Vehicular Technology, vol. 70, no. 5, pp. 4037-
4050, May 2021, doi: 10.1109/TVT.2021.3074820.
[5] Y. Saleem et al., "IoTRec: The IoT Recommender for Smart Parking 
System," in IEEE Transactions on Emerging Topics in Computing, 
vol. 10, no. 1, pp. 280-296, 1 Jan.-March 2022, doi: 
10.1109/TETC.2020.3014722.
[6] E. H. -K. Wu, J. Sahoo, C. -Y. Liu, M. -H. Jin and S. -H. Lin, "Agile 
Urban Parking Recommendation Service for Intelligent Vehicular 
Guiding System," in IEEE Intelligent Transportation Systems 
Magazine, vol. 6, no. 1, pp. 35-49, Spring 2014, doi: 
10.1109/MITS.2013.2268549.
[7] C. -L. Chen and W. -C. Chiu, "A recommendation model of smart 
parking," 2017 13th International Conference on Natural 
Computation, Fuzzy Systems and Knowledge Discovery (ICNCFSKD), Guilin, China, 2017, pp. 2762-2766, doi: 
10.1109/FSKD.2017.8393216.
[8] G. Baranwal, D. Kumar and D. P. Vidyarthi, "A Multi-Criteria 
Framework for Smart Parking Recommender System," 2020 IEEE 
International Smart Cities Conference (ISC2), Piscataway, NJ, USA, 
2020, pp. 1-8, doi: 10.1109/ISC251055.2020.9239098.
48
[9] K. S. Liu, J. Gao, X. Wu and S. Lin, "On-Street Parking Guidance 
with Real-Time Sensing Data for Smart Cities," 2018 15th Annual 
IEEE International Conference on Sensing, Communication, and 
Networking (SECON), Hong Kong, China, 2018, pp. 1-9, doi: 
10.1109/SAHCN.2018.8397113.
[10]S. Yan, N. E. O'Connor and M. Liu, "U-Park: A User-Centric Smart 
Parking Recommendation System for Electric Shared Micromobility 
Services," in IEEE Transactions on Artificial Intelligence, doi: 
10.1109/TAI.2024.3428513.
[11]A. Yavari, P. P. Jayaraman and D. Georgakopoulos, "Contextualised 
service delivery in the Internet of Things: Parking recommender for 
smart cities," 2016 IEEE 3rd World Forum on Internet of Things (WFIoT), Reston, VA, USA, 2016, pp. 454-459, doi: 10.1109/WFIoT.2016.7845479.
[12]H. Sun, X. Huang and W. Ma, "Beyond Prediction: On-Street Parking 
Recommendation Using Heterogeneous Graph-Based List-Wise 
Ranking," in IEEE Transactions on Intelligent Transportation 
Systems, vol. 25, no. 6, pp. 5892-5903, June 2024, doi: 
10.1109/TITS.2023.3336808.
[13]LightGCN: Simplifying and Powering Graph Convolution Network 
for Recommendation Xiangnan He, Kuan Deng, Xiang Wang, Yan Li, 
YongDong Zhang, and Meng Wang. 2020. LightGCN: Simplifying 
and Powering Graph Convolution Network for Recommendation. In 
Proceedings of the 43rd International ACM SIGIR Conference on 
Research and Development in Information Retrieval (SIGIR '20). 
Association for Computing Machinery, New York, NY, USA, 639–
648.