現今網路快速發展下，帶動了數位音樂的普及，使用者漸漸能藉助網路隨時聆聽到他們所喜好的音樂歌曲，更在電子商務引領下發展出推薦系統，提高使用者對音樂購買的慾望。目前一般線上音樂推薦系統，經常擷取從過去的歷史紀錄，以進行資料分析或透過統計方式來推薦其它人常聽的歌曲，但往往使用者的需求不僅如此，因為使用者經常受時間或地點等因素，而選擇或改變聽歌的行為，若系統只考慮到使用者喜好的歌曲類型，似乎對個人化的推薦服務仍不夠完整健全。所以本研究於音樂播放清單中欲加入時間排程的概念，並結合決策樹分類技術，期望推薦更符合使用者當前適合的音樂。最後，經由實作與分析後的推薦準確度結果，已達我們所預期的目標成效。

The digital music is booming through rapid expansion of Internet. Users can listen to their favorite songs via web all the time. The electronic commerce leads to development of the recommendation system, which enhances the desire of music buying for customers. The online music recommendation system usually grabs the historical record from past listeners. With extensive data analysis or statistical means, the system recommends the popular music to others. However, users’ demand is far beyond that because their choice or change of listening behaviors is often affected by various factors, such as time or place. If the system only considers the type of favorite songs for users, it seems not a comprehensive service for personal recommendation system. So this research will add time scheduling to the music playlist, and combines classification technology of decision tree to suggest users the suit music more precisely. Eventually, the accuracy of recommended results achieved in our anticipate result after implementation and analysis.

摘要........................................ii
Abstract...................................iii
誌謝.........................................v
目錄.........................................vi
圖索引.......................................ix
表索引.........................................x
1. 緒論.....................................1
1.1 研究動機.................................1
1.2 研究目的.................................2
1.3 論文架構.................................3
2. 文獻探討..................................4
2.1 推薦系統.................................4
2.1.1內容導向式(Content-Based, CB)..............4
2.1.2合作式過濾法(Collaborative Filtering)......6
2.1.3混合式推薦法(Hybrid Recommendation)........9
2.1.4情境推薦法(Hybrid Recommendation).........11
2.2 播放清單產生器(Playlist Generation)......12
2.3 分類學習法(Classification Learning)......13
2.3.1 ID3分類學習法............................14
2.3.2 C4.5分類學習法...........................15
2.4 音樂特徵值擷取............................18
3. 問題定義與方法............................22
3.1 問題定義.................................22
3.2 系統架構與方法............................23
3.3 C4.5決策樹虛擬碼(Virtual Code)............28
3.4 決策樹之建立..............................29
3.5 使用者行為模型的建置.......................32
3.6 產生個人音樂排程清單.......................34
3.7 改進灰姑娘問題之策略.......................35
4. 實驗分析..................................37
4.1 系統環境建立與評估.........................37
4.2 系統運作流程...............................38
4.3 系統畫面..................................39
4.4 實驗結果與分析.............................40
5. 結論......................................48
5.1 未來研究與建議.............................49
參考文獻..........................................51
作者簡介..........................................56

[1] Amazon.com, “Earth’s biggest selection, includes personalized recommendation and wish lists for registered users,” http://www. amazon.com.
[2] Arthur F., Dominik S., Martin G., Gerhard W., “Playlist Generation Using Start and End Songs,” Music Recommendation and Organization, 2008, PP.173-178.
[3] Beth L., “Music recommendation from song sets,” International Conference on Music Information Retrieval, 2004, pp.425–428.
[4] Bamshad M., Honghua D., Tao L., Yuqing S., Jiang Z., “Combining Web Usage and Content Mining for More Effective Personalization,” In Proceedings of the International Conference on E-Commerce and Web Technologies, 2000.
[5] Bello, J. P., Daudet L. and MB S., “Time-domain Polyphonic Transcription using Self-generating Databases,” in Proceeding of the 112th convention of the audio engineering Society, Munich, Germany, 2002.
[6] Balabanovic, M. and Shoham, Y. “Learning Information Retrieval Agents: Experiments with Automated Web Browsing,” In Proceedings of the AAAI Spring Symposium on Information Gathering from Heterogeneous, 1995.
[7] Claudiu S. F., Wolfgang N., Raluca P., “The Benefit of Using Tag-Based Profiles,” In Proceedings of the 2007 Latin American Web Conference (LA-WEB), 2007, PP.32-41.
[8] Chen H. C., Chen Arbee L. P., “A music recommendation system based on music data grouping and user interests,” Proceedings of ACM International Conference on Information and Knowledge Management, 2001.
[9] Daniel M., Sally J. C., “Sociology And Music Recommendation Systems,” Austrian Computer Society(OCG) ,2007.
[10] Elias P., Tim P., Gerhard W., “Dynamic Playlist Generation Based on Skipping Behavior,” Austrian Research Institute for Artificial Intelligence (OFAI), 2005, pp.634-637.
[11] Fu, Y., “Data mining: tasks, techniques, and applications,” IEEE Potentials, Vol.16, No.4, 1997, pp.18-20.
[12] Fayyad, U. M. and Irani, K. B., “On the handling of continuous-valued attributes in decision tree generation,” Machine Learning, 1992, pp.87-102.
[13] Goldberg, D., Nichols D., and Terry D., “Using Collaborative Filtering to Weave an Information Tapestry,” Communications of the ACM, 1992.
[14] Herlocker J. L. and Konstan J. A., “Content-Independent Task Focused Recommendation,” IEEE Internet Computing, 2001, pp.40-47.
[15] Hammer E. R. and Malcolm S. C. “Guided listening; a textbook for music appreciation,” 1992.
[16] Herlocker, J., Konstan, J. A., Borchers, A., and Riedl, J. “An Algorithmic Framework for Performing Collaborative Filtering,” Proceedings of SIGIR 99, Berkeley, pp.230-237, 1999.
[17] iLike.com, “http://www.ilike.com,” Retrieved date: 2007/11/5, 2007.
[18] Konstan J., Miller B., Maltz D., Herlocker J.,Gordon L., and Riedl J., “GroupLens: applying collaborative filtering to use net news., “Communications of the ACM, 1997.
[19] Katsuhiko K., Keiji H., Katashi N., “A Music Recommendation System Based on Annotations about Listeners,” Automated Production of Cross Media Content for Multi-Channel Distribution, 2005.
[20] Kuo F. F. and Shan M. K. “A Personalized Music Filtering System Based on Melody Style Classification,” IEEE International Conference on Multimedia and Expo ICME, 2002, pp.9-12.
[21] Lang K., “News Weeder: Learning to Filter Netnews, “Proceedings of the 12th International Conference on Machine Learning, 1995, pp.331-339.
[22] Lu L., Jiang H. and Zhang H. “A robust audio classification and segmentation method,” ACM Multimedia, 2001, pp.203-211.
[23] Last.fm, “http:///www.last.fm,” Retrieved date: 2007/11/6, 2007.
[24] McKinney M. F., and Breebaart J., “Features for Audio and Music Classification,” In Proc. ISMIR, 2003, pp.151-158.
[25] Monti, G. and Samder, M. “Pitch Locking Monophonic Music Analysis,” in Proceeding of the 112th convention of the audio engineering Society (AES), 2002.
[26] Nuria O., Lucas K. S., “PAPA: Physiology and Purpose-Aware Automatic Playlist Generation“, International Conference on Music Information Retrieval (ISMIR), 2008.
[27] Oscar C., Miquel R., Perfecto Herrera, “A MUSIC RECOMMENDATION SYSTEM BASED ON RSS FEEDS AND USER PREFERENCES,” 6th International Conference on Music Information Retrieval (ISMIR), 2005.
[28] Pandora.com, “http://www.pandora.com,” Retrieved date: 2007/10/20, 2007.
[29] Quinlan, J.R. “C4.5: Programs for Machine Learning,” Morgan-Kaufmann, Morgan Kaufmann Series in Machine Learning, 1993, pp.302.
[30] Raphael, C., “Automatic Transcription of Piano Music,” in Proceedings of the 3rd International Symposium on Music Information Retrieval (ISMIR), 2000.
[31] Rob van G. and Fabio V., “Visual Playlist Generation on the Artist Map,” Institute of Information and Computing Sciences, ISMIR, 2005, pp.520-523.
[32] Slaney M., Weinberger K., White W., “Learning a Metric for Music Similarity, Institute of Information and Computing Sciences, ISMIR, 2008, pp.313-318.
[33] Shan, M. K. Kuo F. F. and Chen M. F. “Music Style Mining and Classification by Melody,” IEEE International Conference on Multimedia and Expo( ICME), Vol. 1, Aug, pp.26-29, 2002.
[34] Steffen P. and Sander van de W., “User evaluation of a new interactive playlist generation concept,” International Conference on Music Information Retrieval, ISMIR, 2005, pp.638-643.
[35] Shardanand U., Maes P., “Social information filtering: algorithms for automating, word of mouth.” In Proceedings of the ACM CHI Conference, Vol. 1, 1995, pp.210-217.
[36] Sarwar B. M., Karypis G., Konstan J. A., and Riedl J., “Application of Dimensionality Reduction in Recommender System-A Case Study,” ACM Web KDD Web Mining for E-Commerce Workshop, 2000.
[37] Srinivasan S., Petkovic D. and Ponceleon D. B., “Towards robust features for classifying audio in the CueVideo system,” In Proceedings of ACM Conf. On Multimedia, 1999, pp.399-407.
[38] Tzeng Y. S., Chen H.C., and Chen A. L. P., “A New Approach for Rating-Based Collaborative Music Recommendation Using Personal Preferences and Opinions From Trustable Users,” Proc. of The IASTED Conference on Internet and Multimedia Systems and Applications(IMSA), 2004.
[39] The Echo Nest Analyze, “API.http://developper.echonest.com/docs/analyze/xml,” xml description, downloaded March 31, 2008.
[40] Vucetic S. and Obradovic Z. “A regression-base approach for scaling-up personalized recommender systems in e-commerce, ACM Web KDD Web Mining for E-Commerce Workshop, 2000.
[41] Wang H., Wang W., Yang J., and Yu P. S., “Clustering by Pattern Similarity in Large Data Sets,” Proceeding of ACM Special Interest Group on Management of Data, 2002.
[42] Yoshinori H., Kazuhiro w., Kazuki T., Shogo N., “Content-based Music Filtering System,” ACM, 2006, pp.1050-1057.
[43] Yu K., Xu X., Ester M., and Kriegel H.P., “Selecting relevant instances for efficient accurate collaborative filtering,” Proceedings of the 10th International conference on Information and knowledge management, 2001, pp.239-246.