使用關聯式分類器於中文意見探勘系統之應用

Application of Chinese Opinion exploration system with Associative Classification

劉鴻郁(Hong-Yu Liu)

指導教授 ： 蔣璿東

淡江大學/工學院/資訊工程學系碩士在職專班/碩士(2018年)

https://doi.org/10.6846/TKU.2018.00922

摘要

相關研究指出，爭取一個新客戶的成本，是維繫一個舊客戶成本的五倍，而網友常會在網路上表達更換或勸人更換原業者的文章。所以本研究將以網際網路服務供應商(Internet Service Provider，ISP)為研究對象，冀望能盡快找出這些文章，爭取時效來處理以減少顧客流失。所以本研究將利用中文意見探勘系統找出可能在表達更換或勸人更換原業者的完整句(opinion sentences)，再使用AC分類器對這些完整句做進一步的分類。但由於原意見探勘系統是針對口碑分析所設計，無法準確表達這類完整句，導致分類效能不佳。本研究在不影響原意見探勘系統功能的前提下，(1)修改系統原完整句表達方式:將含動詞的完整句由四個意見元素的組合改成由七個意見元素的組合而成；(2) 新增回推功能以填補某些完整句所缺的面相資訊。經實驗證明，強化完整句表達的精準度確實可以大幅改善分類的結果。

關鍵字

關聯規則 ； 關聯式分類器 ； 資料探勘

並列摘要

Studies show that gaining one customer costs five times as much as keeping one customer, and web users often use the Internet to express a desire to change providers or encourage others to switch. This study uses Internet Service Providers (ISPs) as the study target, hoping to find these writings as fast as possible in an effort to reduce the loss of custom-ers. This study will make use of Chnese-language opinion mining systems to search for complete sen-tences (opinion sentences) that might be expressing a desire to change providers or encouraging others to switch. It will also use AC classifiers to perform further classification on these sentences. But be-cause the opinion mining systems were designed for analyzing word of mouth reputations, current com-plete sentences cannot accurately this type of textual opinions, resulting in poor classification. On the premise of not affecting the opinion mining system, this study will:
1. .revise the system's sentence conveyance method, changing from four opinion ele-ments to seven opinion elements for sen-tences containing verbs;
2. .introduce a pushback function to fill in exterior information that is lacking in some sentences.
As evidenced by experiments, strengthening the accuracy of complete sentence expression could significantly improve classification results.

並列關鍵字

Association Rules ； Associative Classification ； Data mining

參考文獻

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. 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.

4. 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.

5. 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.