A web-based chatbot tutor was developed to help Grade 8 South African students improve natural science performance. Using SQL, DistilBERT, VADER, and PaLM models, it delivers factual, syllabus-based answers with emotive feedback. The chatbot achieved 85–96% accuracy in topic, intent, and sentiment detection, enhancing learning and engagement.
# Design-of-an-Emotive-Chatbot-Tutor-Focusing-on-Natural-Science-in-South-Africa
A web-based chatbot tutor was developed to help Grade 8 South African students improve natural science performance. Using SQL, DistilBERT, VADER, and PaLM models, it delivers factual, syllabus-based answers with emotive feedback. The chatbot achieved 85–96% accuracy in topic, intent, and sentiment detection, enhancing learning and engagement.