AFAAN OROMOO TEXTUAL ENTAILMENTCLASSIFICATION USINGDEEPLEARNINGAPPROACH

Human communication relies on natural language, such as Afaan Oromoo. However, forcomputers to interact effectively with humans, they must be able to understandnatural language. Natural language processing is the field that enables computers to understandand use human language. Since textual entailment determines the relationship betweensentence pairs, it is a crucial task in natural language processing. Recent development indeep learning has offered promising solutions for automating feature engineeringandlearning semantic representations. This study used a deep learning approach for classifyingtextual entailment in Afaan Oromoo into three categories: Entailment, Contradiction, andNeutral. Despite its widespread use in the Horn of Africa, Natural Language Processingtools for Afaan Oromoo are limited. To address this gap, we collected a dataset of 13,060sentence pairs in Afaan Oromoo, preprocessed the data, and developed model architecturefor classification. We developed the model using various deep learning approaches, including CNN, LSTM, and BiLSTM, comparing their performance to identify themost effective approach. The BiLSTM model showed highest performance, achieving91.23%accuracy on the training dataset, 82.15% accuracy on the validation dataset, and80.47%accuracy on the test dataset. Considering that there are currently little resources availablefor Afaan Oromoo Natural Language Processing, these results are encouraging. Asastarting point for future research, this study offers a basis for additional investigationandadvancement in this field. This research is expected to make a substantial contributiontothe development of Afaan Oromoo's natural language processing capabilities.