微網誌短文分類應用於房價指數之相關性分析

Applying Microblogging Short Text Classification in Correlation Analysis of Housing Price Index

曾暐峰(Wei-Feng Tseng)

指導教授 ： 王正豪

國立臺北科技大學/電資學院/資訊工程系研究所/碩士(2014年)

https://doi.org/10.6841/NTUT.2014.00323

摘要

隨著網路的蓬勃發展，社群網站也逐漸流行，許多使用者會在上面發表自己的意見，對於各領域的資訊也都大量分享以及討論，其中以股價及房價波動大，所以我們探討是否可以以網路上的意見來加以分類，並且觀察其與房價指數的相關性，來做為使用者的參考。
因此，本論文提出一個收集微網誌與房價相關的短文來進行分類，透過短文進行斷詞及特徵擷取來判斷這短文是評論房價漲或跌，並以相關係數分析其與實際上真正的房價相比的相關性，以微網誌意見提供實際房價波動的參考。
本實驗在PLURK上收集15個月的房價相關短文來分類，每月大概有300~500筆資料，實驗結果顯示，與各縣市實際房價指數的相關係數最高達到0.8，表示為有相當高的相關性，並驗證了所提方法能有效反映實際房價的漲跌。

關鍵字

短文分類 ； 特徵擷取 ； 相關係數

並列摘要

With the rapid development of the internet, social networking sites are becoming more popular, many users express their views in the various fields. With large fluctuations in stock index and house prices, we explore to analyze user opinions by short text classification, and observe its correlation with the housing-related price index.
In this paper, we collected the short texts in micro-blog for classification of the sentiment orientation. Then, we analyze its correlation with the real housing prices by correlation coefficient.
In our experiments, we collected 15 months of data on Plurk, in which there are about 300 to 500 messages per month. From the experimental results, the correlation coefficient between the proposed approach and the actual housing price index can be up to 0.8, expressed as a very high correlation. This validates the effectiveness of applying short text classification in microblog for reflecting the fluctuations of real house prices.

並列關鍵字

Document Classification ； Feature Extraction ； correlation coefficient

參考文獻

[18]李啟菁,”中文部落格文章之意見分析”，碩士論文，國立台北科技大學資訊工程研究所，2010。

[2] 陳維君, “網路意見分析與意見調查比較之研究 : 以公共議題為例”,碩士論文,淡江大學資訊管理學系碩士班, 2013

[4] Jeremy C. Ang,“Prosodic Cues for Emotion Recognition in Communicator Dialogs,”M.S thesis, University of California Berkeley, 2002.

[5] David B. Bracewell,“Semi-Automatic Creation of an Emotion Dictionary Using WordNet and its Evaluation,” Proceedings of IEEE conference on Cybernetics and Intelligent System,pp.21-24,2008.

[6] Apoorv Agarwal, Boyi Xie, Ilia Vovsha, Owen Rambow, Rebecca Passonneau, “Sentiment Analysis of Twitter Data”, Proceeding LSM '11 Proceedings of the Workshop on Languages in Social Media Pages 30-38, 2011