利用原始語意元衡量概念之語意關聯性

Measuring Concept Semantic Relatedness Based on Semantic Primitives

許友惠(Hsu, Yu Hui)

指導教授 ： 蘇豐文

國立清華大學/電機資訊學院/資訊系統與應用研究所/碩士(2015年)

摘要

近年來，越來越多的科技與自然語言處理有著密不可分的關係。而在自然語言處理中，如何判斷不同字詞之間的相關性是很重要的技術之一。
　　在本篇論文中，應用了一個基於常理建立的知識庫並且找出其中的原始語意；接著提出方法來衡量兩個字詞之間的相關性。首先，我們利用隨機漫步演算法來分析這個知識庫，接著利用HITS 演算法(一種常見的網頁排名演算法)找出知識庫中的原始語意。接著，我們提出了兩個演算法來衡量兩個字詞之間的相關性。
　　最後我們計算與標準答案的斯皮爾曼等級相關係數(Spearman’s correlation)，得到0.54~0.8的結果。

關鍵字

語意相關性分析 ； 原始語意元 ； 自然語言處理

並列摘要

Measuring semantic relatedness is one of the important fundamental technical processes. In this thesis, we propose an approach to find the semantic primitives embedded in a common sense database (ConceptNet) and the algorithms to measure the concept semantic relatedness. We used the Random Walk Algorithm to analyze the common sense database first, and adopt the HITS, a well-known web rank algorithm, to find the semantic primitives in this database. Then we propose two algorithms to measure the semantic relatedness between different pairs of concepts.
We adopted the Spearman’s correlation score as criteria of semantic relatedness and compared the performance of our methods against some benchmark data. Our performance in terms of Spearman’s correlation score ranging from 0.54 to 0.8.

並列關鍵字

Semantic Relatedness Analysis ； Semantic Primitives ； Natural Language Processing

參考文獻

[4] Strube, Michael, and Simone Paolo Ponzetto. "WikiRelate! Computing semantic relatedness using Wikipedia." AAAI. Vol. 6. 2006.

[5] Gabrilovich, Evgeniy, and Shaul Markovitch. "Computing Semantic Relatedness Using Wikipedia-based Explicit Semantic Analysis." IJCAI. Vol. 7. 2007.

[8] Radinsky, Kira, et al. "A word at a time: computing word relatedness using temporal semantic analysis." Proceedings of the 20th international conference on World wide web. ACM, 2011.

[10] HITS Algorithm. Kleinberg, Jon M. "Authoritative sources in a hyperlinked environment." Journal of the ACM (JACM) 46.5 (1999): 604-632.

[11] Finkelstein, Lev, et al. "Placing search in context: The concept revisited."Proceedings of the 10th international conference on World Wide Web. ACM, 2001.