臺灣博碩士論文加值系統

根據英國慈善援助基金會(CAF)的世界奉獻指數報告，全世界有近五分之一的人口於空閒時間會進行志願服務的工作，且每年的人數仍在持續成長當中，這樣的情形也帶動了非營利組織與志工機會數量的增加，使用者要從如此大量的志工活動中找尋到合適的活動就如同大海撈針。本研究嘗試利用隱含狄利克雷分佈模型建構出志工機會、志工類型及使用者三者之間的關聯，並以銀髮族作為範例族群，建立一套志工活動推薦系統，協助使用者在選擇志工活動時進行決策。
本研究設計了一套志工活動推薦系統，除了由使用者過去參加過的志工活動建構出隱含狄利克雷分佈模型中志工機會、志工類型及使用者三者的關聯性之外，並由過去參加過的志工活動推論出使用者時間地點的限制因素，並收集使用者於社群平台中的互動行為，找出親近的使用者及其喜好，進而推薦使用者可能合適並有興趣的志工活動。
實驗評估利用十次交叉驗證法對三組不同資料集分別測試系統效能，比較使用隱含狄利克雷分佈與類別基礎法與傳統協同式過濾的效能差異，由實驗結果得知使用隱含狄利克雷分佈方法準確率達到最高，而類別基礎法與協同式過濾效能不相上下，唯資料量稀疏時，類別基礎法勝過協同式過濾許多。再者我們測試了加上社群互動資訊及時間地點因素後，是否提升系統效能。結果得知加上社群互動資訊時系統效能雖有提升，但並不顯著；而加入時間地點因素後效能則是有顯著地提升。
本研究以銀髮族為範例族群，研發了一套基於隱含狄利克雷分佈之志工活動推薦系統，透過使用者過去所參加過的志工活動及社群平台的互動資訊，建構出使用者的志工推薦模型，在使用者尋找志工活動時協助進行決策，幫助使用者找尋合適且有興趣之志工活動。

According to the World Giving Index published by the Charities Aid Foundation, nearly 1/5 people of the world population do volunteer service during their free time and this number is on the rise every years. This has led to an increase in the number of nonprofit organizations and volunteer opportunities and it is hence becoming increasingly difficult for users interested in volunteering to choose a suitable opportunity from such a massive database. This research tried to use the latent Dirichlet distribution (LDA) model to construct triangular relations among volunteer opportunities, volunteer types, and users. We chose seniors as the target of our case study to build a volunteer opportunity recommendation system, which assists seniors in making decisions when choosing volunteer opportunities.
This research developed an LDA-based volunteer opportunity recommendation system that constructs triangular relations among volunteer opportunities, volunteer types, and users. In addition to LDA, we take temporal and geographical constraints and social behavior in a senior-focused online social network into account to assist seniors in decision making when choosing volunteer opportunities.
For the experiment evaluation, we used ten-fold cross-validation to test and evaluate recommendation models in three different dataset we collected. We compared the three recommendation methods to observe their performance, namely, LDA model, category-based, and traditional collaborative filtering. According to experimental results, LDA model had the best performance in accuracy, and category-based and collaborative filtering were neck and neck. But only for sparse dataset, category-based outperformed collaborative filtering. Furthermore, we investigated whether our method improved the performance of recommendation models by considering social information and temporal and geographical constraints. Results showed that when considering temporal and geographical constraints, the performance significantly improved, which indicated that temporal and geographical constraints were important factors when recommending volunteer opportunities to users. However, social information slightly improved performance, but the effect was not significant.
This research targeted seniors for our case study and developed a LDA-based volunteer opportunity recommendation system which used historical volunteer opportunities which the user had participated in the past and social behavior on the online social network to build recommendation models to assist seniors in decision making when choosing volunteer opportunities.

中文摘要 I
Abstract III
Contents VI
List of Tables VIII
List of Figures IX
Chapter 1 Introduction 1
1.1 Background 1
1.2 Research Objective and Specific Aims 3
1.3 Thesis Organization 3
Chapter 2 Related Work 4
2.1 Volunteer Matching Platform and Recommendation System 4
2.2 Overview of Recommendation Systems 5
2.2.1 Collaborative Filtering 5
2.2.2 Content-based Filtering 6
2.2.3 Hybrid Methods 7
2.3 Trust-based Recommendation System 7
2.4 LDA Model and Application in Recommendation System 7
Chapter 3 Materials and Methods 9
3.1 Data Pre-processing 10
3.2 User Condition 12
3.3 User Similarity 14
3.3.1 Category-based Method 15
3.3.2 LDA Model 15
3.4 Acquainted User 17
3.5 Trusted User 18
3.6 Volunteer Opportunity Score 18
Chapter 4 Experiments 20
4.1 Experimental Design 20
4.2 Data Collection 21
4.2.1 Volunteer Record 21
4.2.2 Interaction Pattern 22
4.3 Evaluation 23
4.4 Experimental Results 23
4.4.1 Dataset Description 24
4.4.2 Similar User Selection 26
4.4.3 Social information and User Condition 28
Chapter 5 Conclusions and Future Work 30
5.1 Conclusions 30
5.2 Future Work 31
References 32

Agarwal, D. and Chen, B.-C., fLDA: matrix factorization through latent dirichlet allocation, In: Proceedings of the third ACM international conference on Web search and data mining, New York, New York, USA: ACM, 91-100, 2010.
Blei, D. M., Ng, A. Y. and Jordan, M. I., Latent dirichlet allocation, J. Mach. Learn. Res., 3: 993-1022, 2003.
Chen, W.-C., Cheng, Y.-M., Sandnes, F. E., et al., Finding Suitable Candidates: The Design of a Mobile Volunteering Matching System, In: Jacko, J.A., editor HCI (3), Springer, 21-29, 2011.
Deshpande, M. and Karypis, G., Item-based top-N recommendation algorithms, ACM Trans. Inf. Syst., 22: 143-177, 2004.
Foundation, C. A., The World Giving, 2014.
Herlocker, J. L., Konstan, J. A., Terveen, L. G., et al., Evaluating Collaborative Filtering Recommender Systems, ACM Trans. Inf. Syst., 22: 5--53, 2004.
Kane, S. K. and Klasnja, P. V., Supporting volunteer activities with mobile social software, In: CHI '09 Extended Abstracts on Human Factors in Computing Systems, Boston, MA, USA: ACM, 4567-4572, 2009.
Krestel, R., Fankhauser, P. and Nejdl, W., Latent dirichlet allocation for tag recommendation, In: Proceedings of the third ACM conference on Recommender systems, New York, New York, USA: ACM, 61-68, 2009.
Lee, D. H. and Brusilovsky, P., Social networks and interest similarity: the case of CiteULike, In: HT '10: Proceedings of the 21st ACM Conference on Hypertext and Hypermedia, New York, NY, USA: ACM, 151-156, 2010.
Liu, Q., Chen, E., Xiong, H., et al., Enhancing Collaborative Filtering by User Interest Expansion via Personalized Ranking, IEEE Transactions on Systems, Man, and Cybernetics, Part B, 42: 218-233, 2012.
Ma, H., King, I. and Lyu, M. R., Effective missing data prediction for collaborative filtering, In: Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval, Amsterdam, The Netherlands: ACM, 39-46, 2007.
Melville, P., Mooney, R. J. and Nagarajan, R., Content-Boosted Collaborative Filtering for Improved Recommendations, In: Dechter, R. and Sutton, R.S., editors, AAAI/IAAI, AAAI Press / The MIT Press, 187-192, 2002.
Mooney, R. J. and Roy, L., Content-based book recommending using learning for text categorization, In: Proceedings of the fifth ACM conference on Digital libraries, San Antonio, Texas, USA: ACM, 195-204, 2000.
Morse, J., Cerretani, J., Halai, S., et al., doGooder: fostering volunteer communities to serve the homeless, In: Czerwinski, M., Lund, A.M. and Tan, D.S., editors, CHI Extended Abstracts, ACM, 3843-3848, 2008.
Sarwar, B., Karypis, G., Konstan, J., et al., Item-based collaborative filtering recommendation algorithms, In: Proceedings of the 10th international conference on World Wide Web, Hong Kong, Hong Kong: ACM, 285-295, 2001.
Siehndel, P., Kawase, R., Hadgu, A. T., et al., Finding relevant missing references in learning courses, In: Proceedings of the 22nd international conference on World Wide Web companion, Rio de Janeiro, Brazil: International World Wide Web Conferences Steering Committee, 425-430, 2013.
VolunteerMatch, https://www.volunteermatch.org/, 2015.
Wang, C. and Blei, D. M., Collaborative topic modeling for recommending scientific articles, In: Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining, San Diego, California, USA: ACM, 448-456, 2011.
Ziegler, C.-N. and Golbeck, J., Investigating interactions of trust and interest similarity, Decision Support Systems, 43: 460 - 475, 2007.
Ziegler, C.-N. and Lausen, G., Analyzing Correlation between Trust and User Similarity in Online Communities, In: Jensen, C.D., Poslad, S. and Dimitrakos, T., editors, iTrust, Springer, 251-265, 2004.