Abstract
Analysis of bibliographic data and query logs in the Japanese Agricultural Sciences Index showed that keywords in user queries did not generate hits for all relevant documents by synonyms for query terms. Both the notation system (kanji, katakana, and hiragana) used to write crop names and the Japanese terms used to name proteins differed among papers published by different academic societies. For example, 63% of users selected only one Japanese name for sweet potato as a query term, potentially missing documents that used other Japanese names for it or getting documents irrelevant to their need. Our findings suggest that the provision of synonyms for query terms and of the number of scientific papers which include each synonym will help users to maximize coverage and get documents relevant to their need.
