隨著網路技術的快速發展，大部份消費者在購買某項產品或某公司的服務時，會先瀏覽網站上的相關評論之後，再進行購買行為。相關研究指出，評論文章會影響消費者對產品或某公司服務的購買決策。所以對於公司而言，公司利用顧客評論的文章做各面向(aspect)的口碑分析及查看網友們的意見，以便於能在最短的時間內回覆和平衡與公司相關的負面評價，這是件非常重要的工作。因此就台灣的公司必須利用中文意見探勘系統來做各面向的口碑分析，我們已經初步發展了一個屬於aspect-level的中文意見探勘系統；此系統分別利用default topic和 default feature來增加意見和所要討論面向的回收率。此研究我們將利用意見元素間的上下文關聯性讓系統不但能推論出部分網友在回文中所要討論的topic、面向和子面向，同時亦能修正部分由default topic和 default feature所造成的錯誤；此增加的功能是能讓使用者在做口碑分析時，能獲得更詳細和正確的資訊。

With the development of network technology, before people buy a product or a company's service, most consumers will search the related comments on the social networking sites. Related studies indicate that the product evaluation article will influence the purchase decisions of consumers. Therefore, analyzing WOM (word of mouth) to different aspects and replying negative WOM are important things to companies. Since opinion analysis at document level and sentence level is too coarse to determine users’ opinions precisely, we have developed an aspect-level Chinese opinion mining system for a specific domain. The system uses default topic and default feature to increase recall rate of opinion mining results and aspects, respectively. In this project, we will use context dependent not only to derive some unknown topics, features and sub-features but also to correct some default topic and default feature errors. Consequently, uses can get more detail and correct information from WOM analysis.

目錄
第一章  緒論	1
1-1	研究動機與目的	1
1-2  研究架構	6
第二章  文獻探討	7
2-1  中文意見探勘系統相關研究	7
2-2  本研究室開發的中文意見探勘系統簡介	11
第三章  問題陳述	16
3-1  回文沒有提對象的問題	16
3-2  預設意見元素的問題	18
3-2-1  Default Topic	18
3-2-2  Default Feature 和Default Item	20
第四章  研究方法	24
4-1  意見元素間的上下文關聯性	24
4-2  CDA演算法	28
第五章  實驗結果	32
5-1資料來源	32
5-2 CDA演算法的單月成效分析	36
5-2-1  修正default topic的單月成效	36
5-2-2  修正default feature的單月成效	39
5-2-3  補充面向與子面向資訊的單月成效	41
5-2-4  各面向單月情形	42
5-3 CDA演算法的成效分析	44
5-3-1  資料來源	44
5-3-2  成效分析	45
第六章 結論與未來展望	49
參考文獻	50
附錄A 英文論文	53

圖目錄
圖 1中文意見探勘系統架構圖	12
圖 2 文章下載與配置流程圖	13
圖 3中文意見探勘系統文章分析流程圖	15
圖 4  PTT論壇標題範例(一)	17
圖 5  PTT論壇標題範例(三)	20
圖 6  PTT論壇標題範例(四)	21
圖 7  PTT論壇標題範例(五)	22
圖 8  MOBILE01論壇標題範例(六)	23
圖 9  THE CONTEXT DEPENDENT ALGORITHM	28
圖 10資料來源頻道比例圖	33
圖 11 中華各面向原始評價圖	43
圖 12 經演算法修正後中華各面向評價圖	43

表目錄
表 1意見元素定義表	2
表 2 關鍵字與討論標題對照表	35
表 3  ISP領域2016-01-2016-06月分析資料數量	35
表 4  2016-01月針對DEFAULT TOPIC完整句統計表	38
表 5  DEFAULT TOPIC錯誤完整句統計表	38
表 6  2016-01月針對DEFAULT FEATURE完整句統計表	40
表 7  2016-01月針對增加面向的完整句統計表	41
表 8  2016-01月針對增加子面向的完整句統計表	42
表 9  ISP領域2016-01-2016-06月分析輸出完整句比較表	45
表 10  2016-01~06月DEFAULT TOPIC完整句比較表	46
表 11  2016-01~06月DEFAULT FEATURE完整句統計表	47
表 12  2016-01~06月補充面向資訊統計表	48
表 13  2016-01~06月補充子面向資訊統計表	48

[1]	Z. Li, M. Zhang, S. Ma, B. Zhou, and Y. Sun, "Automatic Extraction for Product Feature Words from Comments on the Web Information Retrieval Technology." vol. 5839, G. Lee, D. Song, C.-Y. Lin, A. Aizawa, K. Kuriyama, M. Yoshioka, et al., Eds., ed: Springer Berlin / Heidelberg, 2009, pp. 112-123.
[2]	C. Zhang, D. Zeng, J. Li, F.-Y. Wang, and W. Zuo, "Sentiment analysis of Chinese documents: From sentence to document level," J. Am. Soc. Inf. Sci. Technol., vol. 60, pp. 2474-2487, 2009.
[3]	M. Chen and T. Yao, "Combining dependency parsing with shallow semantic analysis for Chinese opinion-element relation identification," in Universal Communication Symposium (IUCS), 2010 4th International (pp. 299-305), 2010, pp. 299-305.
[4]	P. Ting-Chun and S. Chia-Chun, "Using Chinese part-of-speech patterns for sentiment phrase identification and opinion extraction in user generated reviews," in Digital Information Management (ICDIM), 2010 Fifth International Conference on (pp. 120-127), 2010, pp. 120-127.
[5]	V. Hatzivassiloglou and K. R. McKeown, "Predicting the semantic orientation of adjectives," 1997, pp. 174-181.
[6]	G. Qiu, B. Liu, J. Bu, and C. Chen, "Expanding domain sentiment lexicon through double propagation," presented at the Proceedings of the 21st international jont conference on Artifical intelligence (pp.1199-1204), Pasadena, California, USA, 2009.
[7]	M. Hu and B. Liu, "Mining and summarizing customer reviews," presented at the Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 168-177), Seattle, WA, USA, 2004.
[8]	S.-M. Kim and E. Hovy, "Determining the sentiment of opinions," presented at the Proceedings of the 20th international conference on Computational Linguistics (pp. 1367), Geneva, Switzerland, 2004.
[9]	S. Bin and C. Kuiyu, "Mining Chinese Reviews," in Data Mining Workshops, 2006. ICDM Workshops 2006. Sixth IEEE International Conference on (pp. 585-589), 2006, pp. 585-589.
[10]	Y. Qiang, S. Wen, and L. Yijun, "Sentiment Classification for Movie Reviews in Chinese by Improved Semantic Oriented Approach," in System Sciences, 2006. HICSS '06. Proceedings of the 39th Annual Hawaii International Conference on (pp. 53b-53b), 2006, pp. 53b-53b.
[11]	L. Zhuang, F. Jing, and X.-Y. Zhu, "Movie review mining and summarization," presented at the Proceedings of the 15th ACM international conference on Information and knowledge management, Arlington, Virginia, USA, 2006.
[12]	X. Ding, B. Liu, and P. S. Yu, "A holistic lexicon-based approach to opinion mining," presented at the Proceedings of the 2008 International Conference on Web Search and Data Mining (pp. 231-240), Palo Alto, California, USA, 2008.
[13]	L. W. Ku, I. C. Liu, C. Y. Lee, K. Chen, and H. H. Chen, "Sentence-Level Opinion Analysis by CopeOpi in NTCIR-7," in Proceedings of the 7th NTCIR Workshop Meeting on Evaluation of Information Access Technologies: Information Retrieval, Question Answering, and Cross-Lingual Information Access (pp. 260-267), Tokyo, Japan, 2008, pp. 260-267.
[14]	S. Morinaga, K. Yamanishi, K. Tateishi, and T. Fukushima, "Mining product reputations on the Web," presented at the Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 341-349), Edmonton, Alberta, Canada, 2002.
[15]	G. A. Miller. (1980). WordNet. Available: http://wordnet.princeton.edu/
[16]	P. J. Stone, D. C. Dunphy, and M. S. Smith, "The General Inquirer: A Computer Approach to Content Analysis," 1966.
[17]	A. Esuli and F. Sebastiani, "Sentiwordnet: A publicly available lexical resource for opinion mining," 2006, pp. 417-422.
[18]	B. Ohana and B. Tierney, "Sentiment classification of reviews using SentiWordNet," in 9th. IT & T Conference, 2009, p. 13.
[19]	G. Qiu, B. Liu, J. Bu, and C. Chen, "Opinion Word Expansion and Target Extraction through Double Propagation," Computational Linguistics, vol. 37, pp. 9-27, 2011/03/01 2011.
[20]	L.-W. Ku, H.-W. Ho, and H.-H. Chen, "Opinion mining and relationship discovery using CopeOpi opinion analysis system," Journal of the American Society for Information Science and Technology, vol. 60, pp. 1486-1503, 2009.
[21]	L. Chien-Liang, H. Wen-Hoar, L. Chia-Hoang, L. Gen-Chi, and E. Jou, "Movie Rating and Review Summarization in Mobile Environment," Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, vol. 42, pp. 397-407, 2012.
[22]	林憲嘉, "針對單一領域中文意見探勘系統之研究與實作," 淡江大學資訊工程學系資訊工程學系碩士在職專班碩士論文, 2015.
[23]	劉文琇, "結合意見探勘系統 應用於口碑行銷(以寬頻為例)," 淡江大學資訊工程學系資訊工程學系碩士在職專班碩士論文, 2013.