AFAAN OROMO CHATBOT MODELLING FOR WHEAT DISEASE DETECTION AND TREATMENT RECOMMENDATION USING DEEP LEARNING APPROACHES

Chatbot technology is one of the many applications of artificial intelligence. A Chatbot is a computer program that uses natural language to simulate human-like conversations between people and computers or other communication devices. A sophisticated program uses natural language to have conversations with computers. The conversational process of answering questions is the primary use case for chatbots. This study focused on developing an Afaan Oromo based Chatbot modelling using deep learning approaches for wheat disease detection and treatment recommendation. By combining natural language processing with deep learning algorithms, the study modelled Chatbot serves as a vital resource for farmers, delivering timely and accurate information in their native language. The first thing we did was gathered information from Ilu Ababor agricultural offices that more relevant input for prepare dataset to model a Chatbot. We then transformed the data into the JSON data format dataset that contains 34 classes after its labeled. After that, we used natural language processing (NLP) approaches to clean up the provided dataset of inconsistencies, incompleteness, and normalization processes. Text feature generation, or getting our deep learning model to comprehend what we input for training, is the next step after preprocessing our data. The collection and preprocessing of relevant agricultural data, followed by the use of a deep neural network (LSTM) to analyze user queries and provide context-specific responses. We used Softmax for data function activation and for regulation adjustment Dropout. The performance evaluation metrics indicated strong high accuracy rates of 0.9577 (95.77%) while the training loss around 0.2102 (21.02%) at 200 epoch that indicates more significantly trained Chatbot that effectiveness helps farmers for make decisions manage the wheat disease.