隨著網路的普及，店家的銷售模式及消費者的購物模式逐漸改變，許多人購物前會事先在網路論壇等平台查詢產品相關評價後才決定是否購買。網路論壇常見的討論方式為發佈一篇與產品之間比較的文章，而許多網友則會在下方留下自己偏好的產品以及一些對產品的看法。
本論文使用的資料來自於批踢踢實業坊Mobilecomm看板文章，實驗語料選自文章中的網友所撰寫的文章推文。研究目的主要希望針對這些文章推文進行極性分析，透過分析結果進一步對產品進行排名。儘管產品的整體排名是一項重要的指標，但現在越來越多消費者會根據相機功能選購手機，因此研究中同時會對相機功能的部分進行排名。
在計算手機品牌分數上，本研究將建立有向加權圖確立每個手機品牌之間的關係，透過定義比較句和主觀判斷句的方式，以這些句子建立每個產品本身的分數以及有向邊的分數。
最後透過正向句和負向句不同的權重計算方式得出兩種不同結論。當使用正向句分數減去負向句分數時，得出當相機功能排名靠前時，手機品牌的分數不會靠後。當使用正向句分數除以負向句分數時，目前暫時無法判斷手機品牌與相機功能的相關性。

With the popularization of the Internet, the sales model of stores and the shopping model of consumers have gradually changed. Many people will check product-related reviews on online forums and other platforms before making decision. A common discussion method on online forums is to post an article to compare products, and many people will reply some opinions about the products below the article.
This thesis downloaded the review data from PTT Mobilecomm. The experimental corpus is selected from the article tweets written by people in the article. The main purpose of the research is to analyze the sentiment of these articles and tweets, and to further rank products according to the analysis results. Although the overall ranking of products is an important measure, more and more consumers choose mobile phones based on camera functions now. The research will also rank based on the camera functions.
In calculating the scores of mobile phone’s brands, this research will establish a directed weighted graph to establish the relationship between each mobile phone’s brand. By defining comparative sentences and subjective sentences, these sentences are used to establish the vertex score and directed edges score of each product.
Finally, two different conclusions are drawn through the different weight calculation methods of positive and negative sentences. When the weight is calculated by the positive sentence score subtracting the negative sentence score, it is concluded that when the camera function score ranks high, the score of the mobile phone brand will not fall behind. When the weight is calculated by the positive sentence score divided by the negative sentence score, it is currently impossible to determine the correlation between the score of mobile phone brand and the score of camera functions.

Decker, R., & Trusov, M. (2010). Estimating aggregate consumer preferences from online product reviews. International Journal of Research in Marketing, 27(4), 293-307.

Ku, L. W., & Chen, H. H. (2007). Mining opinions from the Web: Beyond relevance retrieval. Journal of the Association for Information Science and Technology, 58(12), 1838-1850

Lafferty, J., McCallum, A., & Pereira, F. C. (2001). Conditional random fields: Probabilistic models for segmenting and labeling sequence data.

Li, S., Zha, Z. J., Ming, Z., Wang, M., Chua, T. S., Guo, J., & Xu, W. (2011, July). Product comparison using comparative relations. In Proceedings of the 34th international ACM SIGIR conference on Research and development in Information Retrieval (pp. 1151-1152).
Miao, Q., Zhang, S., Meng, Y., & Yu, H. (2013, May). Domain-sensitive opinion leader mining from online review communities. In Proceedings of the 22nd International Conference on World Wide Web (pp. 187-188).

Netzer, O., Feldman, R., Goldenberg, J., & Fresko, M. (2012). Mine your own business: Market-structure surveillance through text mining. Marketing Science, 31(3), 521-543.

Page, L., Brin, S., Motwani, R., & Winograd, T. (1999). The PageRank citation ranking: Bringing order to the web. Stanford InfoLab.

Wang, S. M., & Ku, L. W. (2016, May). ANTUSD: A large Chinese sentiment dictionary. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC'16) (pp. 2697-2702).

Zhang, Z., Guo, C., & Goes, P. (2013). Product comparison networks for competitive analysis of online word-of-mouth. ACM Transactions on Management Information Systems (TMIS), 3(4), 1-22.

中研院中文斷詞系統：http://ckipsvr.iis.sinica.edu.tw/