A Machine Learning-Based Mobile Chatbot for Crop Farmers

Abstract

Agriculture remains the basis of the country’s economy, providing the main source of livelihood for most citizenry such as food, employment, income and foreign exchange as well as raw materials for the manufacturing sectors. Despite the great need for economic advancement in crop farming, agriculture seems to be limited in some parts of the country as many people go about in search of white-collar jobs due to lack of adequate information and knowledge on the use of modern farming technologies. The inability of farmers in rural and sub-urban areas to access agricultural knowledge and real-time information on latest farming practices to enhance informed decision making related to soil properties, seeds, fertilizers, pests, modern agricultural tools, and agro-best practices leads to poor crop productivity by farmers. This work is aimed at providing a mobile chatbot for crop farmers in Uyo and its environs. The dataset used was obtained from Akwa Ibom State Ministry of Agriculture and farmers using a combination of two classic research methods; questionnaires and interviews. An ontology-based representation of the obtained dataset is used for training the chatbot using a hybridized machine approach that consists of word shuffling and Jacquard Similarity algorithm. The resulting chatbot is a knowledge base that will provide the means of obtaining useful answers to questions, advice and recommendations on specific farming concerns. The use of the chatbot will give government a platform to reach out to famers in the state and obtain feedback on governance through agricultural services.