NLP Models to Clarify Degree Regulations in Peruvian Universities

Apaza H.; Balcona J.E.; Maquera S.M.; Collanqui E.A.; Reynoso A.I.; Coaguila M.E. · IEEE Andescon, ANDESCON 2024 - Proceedings (2024) · DOI: 10.1109/andescon61840.2024.10755580

Resumen

This work examines the use of Natural Language Models (NLP) to answer questions from university graduates about the regulations for obtaining academic degrees and profes-sional titles. The increasing number of inquiries from graduates highlights the complexity and variety of the regulations. The creation of an NLP-based Chatbot is proposed to manage these queries, taking advantage of the ability of NLP models to understand and answer queries in natural language. Various NLP models, including those based on LSTM and GRU, are reviewed, highlighting their effectiveness in semantic understanding and answering specific questions. Of the models, the highest accuracy is BiGRU with 99.57%, followed by BiLSTM, which achieved 99.15% accuracy. In third place is the classic LSTM model with 98.36%, and the classic GRU with 98.10% accuracy. It is argued that chatbots are valuable tools to improve interaction with students and consumers, aligning with the preferences of the target audience.

Instituciones: Universidad Nacional de Moquegua