Semantic Similarity Measures for Malay-English Ambiguous Words
Nurul Husna Mahadzir, Mohd Faizal Omar, Mohd Nasrun Mohd Nawi · Journal of Telecommunication Electronic and Computer Engineering (JTEC) · 2018
The measurement of semantic similarity between words or concepts has been widely applied in various applications such as Artificial Intelligence, Information Processing and Natural Language Processing (NLP) field. In this paper, we are going to demonstrate how several semantic similarity measures can be adapted to disambiguate words that belong to two different languages. The measures include two path-based and three information content (IC)-based measures. These five measures are evaluated against a test bed of 40 word pairs with eight Malay-English ambiguous words. The experimental results on a common benchmark, created by human judgments show that Wu and Palmer's measures have given the best performance as compared to the other four semantic similarity measures.