Cross Language Information Retrieval (CLIR) systems are pivotal in overcoming language barriers for effective information access. This article presents a Malay-Japanese CLIR dictionary designed to facilitate efficient retrieval of information across these languages, focused on the syllabus of Special Preparatory Program to Japan (RPKJ) at University Malaya. This study aims to assist non-native Japanese-speaking students, by developing a dictionary that corresponds to the program’s syllabus, by retrieving relevant Japanese comprehension materials using Malay search query. The primary stages in this system are categorized into two phases, 1) Query Translation and 2) Document Retrieval. Furthermore, the learning resources utilized are “Minna no Nihongo I” and “Minna no Nihongo II” along with a specialized vocabulary list. This system translates Malay words with three Japanese strings: Hiragana, Katakana, and Kanji, and different word types: a total of approximately 2000 nouns and verbs are used as well. Evaluation results demonstrate the effectiveness of the proposed CLIR dictionary in enhancing retrieval performance with average recall of 0.867 and average precision of 0.87 achieved on a test dataset of diverse queries and the prototype effectively managed translations in all three writing systems and effectively retrieved the suitable comprehension documents for the input queries either noun or verb. This shows that the approach is robust for different types of words within the controlled vocabulary. Future work aims to extend this framework to broader language combinations and domains to further enhance cross-cultural information access.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Evaluating Effectiveness of Malay-Japanese Cross Language Information Retrieval (CLIR) Dictionary

  • Fazrina Said,
  • Shaiful Bakhtiar bin Rodzman

摘要

Cross Language Information Retrieval (CLIR) systems are pivotal in overcoming language barriers for effective information access. This article presents a Malay-Japanese CLIR dictionary designed to facilitate efficient retrieval of information across these languages, focused on the syllabus of Special Preparatory Program to Japan (RPKJ) at University Malaya. This study aims to assist non-native Japanese-speaking students, by developing a dictionary that corresponds to the program’s syllabus, by retrieving relevant Japanese comprehension materials using Malay search query. The primary stages in this system are categorized into two phases, 1) Query Translation and 2) Document Retrieval. Furthermore, the learning resources utilized are “Minna no Nihongo I” and “Minna no Nihongo II” along with a specialized vocabulary list. This system translates Malay words with three Japanese strings: Hiragana, Katakana, and Kanji, and different word types: a total of approximately 2000 nouns and verbs are used as well. Evaluation results demonstrate the effectiveness of the proposed CLIR dictionary in enhancing retrieval performance with average recall of 0.867 and average precision of 0.87 achieved on a test dataset of diverse queries and the prototype effectively managed translations in all three writing systems and effectively retrieved the suitable comprehension documents for the input queries either noun or verb. This shows that the approach is robust for different types of words within the controlled vocabulary. Future work aims to extend this framework to broader language combinations and domains to further enhance cross-cultural information access.