Scholarlink

Compound words and multi-word expressions (MWE) are like the building blocks of writing, making it sound nice and following the rules of language. In Hindi (H) and Awadhi (A), these are called ‘Samaas,’ helping to shorten words and make them sound better. The authors developed the algorithm named ABD2KT-1.0, which is good at recognizing six types of ‘Samaas’—‘Avyayibhav’ (AVY), ‘Bahuvrihi’ (BAH), ‘Dvandva’ (DVA), ‘Dvigu’ (DVI), ‘Karmadharya’ (KAR), and ‘Tatpurusha’ (TAT) ‘Samaas.’ These six types of ‘Samaas’ are combinedly known as ‘Samaas’ Category Identifier (SCI). The ABD2KT-1.0 works well for Devanagari languages (DL) and Non-Devanagari languages (NDL). This tool divides the words into two categories—Agreed word and Disagreed word. The word following the rules is known as Agreed word, and the word not following the rules is known as Disagreed word. This research paper focuses on developing a comprehensive list of words from dozens of languages. The 84.1% accuracy is achieved for Awadhi and Hindi languages from ABD2KT-1.0. The other result says that ‘Samaas’ Finder is successfully identifying BAH SCI but facing problems with TAT SCI.

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

Identification of Multi-word Expressions Using Language-Specific Composite Master Lists

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Hema Gaikwad,
  • Jatinderkumar R. Saini

摘要

Compound words and multi-word expressions (MWE) are like the building blocks of writing, making it sound nice and following the rules of language. In Hindi (H) and Awadhi (A), these are called ‘Samaas,’ helping to shorten words and make them sound better. The authors developed the algorithm named ABD2KT-1.0, which is good at recognizing six types of ‘Samaas’—‘Avyayibhav’ (AVY), ‘Bahuvrihi’ (BAH), ‘Dvandva’ (DVA), ‘Dvigu’ (DVI), ‘Karmadharya’ (KAR), and ‘Tatpurusha’ (TAT) ‘Samaas.’ These six types of ‘Samaas’ are combinedly known as ‘Samaas’ Category Identifier (SCI). The ABD2KT-1.0 works well for Devanagari languages (DL) and Non-Devanagari languages (NDL). This tool divides the words into two categories—Agreed word and Disagreed word. The word following the rules is known as Agreed word, and the word not following the rules is known as Disagreed word. This research paper focuses on developing a comprehensive list of words from dozens of languages. The 84.1% accuracy is achieved for Awadhi and Hindi languages from ABD2KT-1.0. The other result says that ‘Samaas’ Finder is successfully identifying BAH SCI but facing problems with TAT SCI.