Sistem Chatbot Layanan Informasi pada Klinik Hewan Menggunakan Glove Embedding dan Algoritma Long Short-Term Memory (LSTM)

Abstract

Veterinary clinics serve as a place of treatment, a source of information and education on appropriate care for pets. The lack of information services in veterinary clinics can lead to uncertainty about a pet's health condition and hinder efforts to provide optimal care. This makes it necessary to have an information service that can be accessed at any time and place. Chatbot is one of the question and answer platforms that can help the information distribution process. The chatbot system using GloVe Embedding and Long Short-Term Memory algorithm can provide answers to user questions related to clinic services and general pet health information. After getting the requested information, users can schedule an appointment to the clinic. The data used comes from local veterinary clinics and a pet health information site called PetCoach, with answers that have been validated directly by veterinarians. System tests conducted to measure the chatbot's performance show that the chatbot has an accuracy rate of 94% with an average response time of around 1.31 seconds.