Comparable Entity Mining from Comparative Questions

Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, pages 650–658, Uppsala, Sweden, 11-16 July 2010. c⃝2010 Association for Computational Linguistics Comparable Entity Mining from Comparative Questions Shasha Li1，Chin-Yew Lin2，Young-In Song2，Zhoujun Li3 1National University of Defense Technology, Changsha, China 2Microsoft Research Asia, Beijing, China 3Beihang University, Beijing, China shashali@nudt.edu.cn1, {cyl,yosong}@microsoft.com2, lizj@buaa.edu.cn3 Abstract Comparing one thing with another is a typical part of human decision making process. How- ever, it is not always easy to know what to compare and what are the alternatives. To ad- dress this difficulty, we present a novel way to automatically mine comparable entities from comparative questions that users posted on- line. To ensure high precision and high recall, we develop a weakly-supervised bootstrapping method for comparative question identification and comparable entity extraction by leveraging a large online question archive. The experi- mental results show our method achieves F1- measure of 82.5% in comparative question identification and 83.3% in comparable entity extraction. Both significantly outperform an existing state-of-the-art method. 1 Introduction Comparing alternative options is one essential step in decision-making that we carry out every day. For example, if someone is interested in cer- tain products such as digital cameras, he or she would want to know what the alternatives are and compare different cameras before making a purchase. This type of comparison activity is very common in our daily life but requires high knowledge skill. Magazines such as Consumer Reports and PC Magazine and online media such as CNet.com strive in providing editorial com- parison content and surveys to satisfy this need. In the World Wide Web era, a comparison ac- tivity typically involves: search for relevant web pages containing information about the targeted products, find competing products, read reviews, and identify pros and cons. In this paper, we fo- cus on finding a set of comparable entities given a user‟s input entity. For example, given an enti- ty, Nokia N95 (a cellphone), we want to find comparable entities such as Nokia N82, iPhone and so on. In general, it is difficult to decide if two enti- ties are comparable or not since people do com- pare apples and oranges for various reasons. For example, “Ford” and “BMW” might be compa- rable as “car manufacturers” or as “market seg- ments that their products are targeting”, but we rarely see people comparing “Ford Focus” (car model) and “BMW 328i”. Things also get more complicated when an entity has several functio- nalities. For example, one might compare “iPhone” and “PSP” as “portable game player” while compare “iPhone” and “Nokia N95” as “mobile phone”. Fortunately, plenty of compara- tive questions are posted online, which provide evidences for what people want to compare, e.g. “Which to buy, iPod or iPhone?”. We call “iPod” and “iPhone” in this example as comparators. In this paper, we define comparative questions and comparators as:  Comparative question: A question that in- tends to compare two or more entities and it has to mention these entities explicitly in the question.  …

Content truncated to 500 words