Term Aggregation: Mining Synonymous Expressions using Personal Stylistic Variations

Term Aggregation: Mining Synonymous Expressions using Personal Stylistic Variations Akiko Murakami Tetsuya Nasukawa IBM Research, Tokyo Research Laboratory 1623-14 Shimotsuruma, Yamato, Kanagawa 242-8502, Japan �akikom, nasukawa �@jp.ibm.com Abstract We present a text mining method for finding synonymous expressions based on the distribu- tional hypothesis in a set of coherent corpora. This paper proposes a new methodology to im- prove the accuracy of a term aggregationsystem using each author’s text as a coherent corpus. Our approach is based on the idea that one per- son tends to use one expression for one mean- ing. According to our assumption, most of the words with similar context features in each au- thor’s corpus tend not to be synonymousexpres- sions. Our proposed method improves the accu- racy of our term aggregation system, showing that our approach is successful. 1 Introduction The replacement of words with a representative synonymous expression dramatically enhances text analysis systems. We developed a text mining sys- tem called TAKMI (Nasukawa, 2001) which can find valuable patterns and rules in text that indicate trends and significant features about specific top- ics using not only word frequency but also using predicate-argument pairs that indicate dependencies among terms. The dependency information helps to distinguish between sentences by their meaning. Here are some examples of sentences from a PC call center’s logs, along with the extracted dependency pairs: � customer broke a tp � customer...break, break...tp � end user broke a ThinkPad � end user...break, break...ThinkPad In these examples, “customer” and “end user” and “tp” and “ThinkPad” can be assumed to have the same meaning in terms of this analysis for the call center’s operations. Thus, these two sentences have the same meaning, but the differences in ex- pressions prevent us from recognizing their iden- tity. The variety of synonymous expressions causes a lack of consistency in expressions. Other exam- ples of synonymous expressions are: customer = cu = cus = cust = end user = user = eu Windows95 = Win95 = w95 One way to address this problem is by assign- ing canonical forms to synonymous expressions and variations of inconsistent expressions. The goal of this paper is to find those of synonymous expres- sions and variations of inconsistent expressions that can be replaced with a canonical form for text analy- sis. We call this operation “term aggregation”. Term aggregation is different from general synonym find- ing. For instance, “customer” and “end user” may not be synonyms in general, but we recognize these words as “customer” in the context of a manufac- turers’ call center logs. Thus, the words we want to aggregate may not be synonyms, but their role in the sentences are the same in the target domain from the mining perspective. Yet, we can perform term aggregation using the same methods as in synonym finding, such as using word feature similarities. There are several approaches for the automatic extraction of synonymous expressions, such as us- ing word context features, but the results of such approaches tend to contain some …

Content truncated to 500 words