Deep Learning-Based Idiomatic Expression Recognition for the Amharic Language

Abstract Idioms are used in Amharic to conceal information or to express ideas indirectly. However, most natural language processing models used with the Amharic language, such as machine translation, semantic analysis, sentiment analysis, information retrieval, question answering, and next word prediction, do not consider idiomatic expressions. As a result, in this paper, we proposed a conventional neural network (CNN) with a FastText embedding model for detecting idioms in an Amharic text. We collected 1700 idiomatic and 1600 non-idiomatic clause datasets from Amharic books to test the proposed model's performance. The proposed model is then evaluated using this dataset. With testing and training datasets, the proposed model achieves an accuracy of 80% and 98%, respectively. We compared the proposed model to other machine learning models like K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Random Forest classifiers. According to the experimental results, the proposed model produces promising results.