Developing Afaan Oromo Text Based Chatbot for Traditional Food Recipe Using Machine Learning

Abstract

Cooking traditional foods, such as Chechebsa and Cuukkoo, presents challenges for both beginners and experienced cooks due to the intricate details and nuances inherent in traditional recipes. To address this, our study proposes the development of a text-based chatbot for assisting individuals in preparing traditional Afaan Oromo recipes using machine learning techniques. The primary problem addressed by this research is the lack of accessible and user-friendly resources for individuals seeking guidance and assistance in cooking traditional dishes in Afaan Oromo. Our aim is to design and implement text-based chatbot capable of understanding user queries, providing step-by-step cooking instructions, and offering personalized cooking advice. The methodology employed in this study involves systematic data collection from various reliable sources, annotation, and curation of a representative dataset of Afaan Oromo food recipes, followed by extensive data preprocessing and model building using machine learning techniques. Experimental scenarios include training and evaluating the chatbot model on the curated dataset to assess its effectiveness, accuracy, and user satisfaction. Key findings indicate that the developed chatbot model exhibits promising performance, achieving a test accuracy of 80%, in assisting individuals with various levels of cooking experience in preparing traditional Afaan Oromo recipes. The implications for the future involve potential applications of the chatbot in culinary education, cultural preservation, and promoting awareness and appreciation of Afaan Oromo cuisine and cultural heritage. Overall, this study contributes to the advancement of natural language processing and machine learning research while addressing practical challenges in the domain of traditional food preparation and cultural preservation.