<![CDATA[XYZ organization develops a website to help the undergraduate students and alumni in UPN “Veteran” Jawa Timur keeping their pace for seeking proper career. However, it is hard for XYZ organization to answer the questions regarding career consultation without having an automation system that can be available for 24 hours. This is why XYZ organization intends to implement a chatbot inside of the website. The research methods for implementing it are based on CRISP-DM model with 1 process iteration and the use of Deep Feed-Forward Neural Network architecture along with TensorFlow.js library for running the chatbot. The result of the research is a chatbot successfully implemented inside of XYZ organization website having the accuracy 99.58%, precision 99.78%, recall 99.70%, and f1-score 99.73%. The total speed for accessing the chatbot is 0.859 second and total chatbot size is 385 kilobyte. Despite that, the chatbot knowledge is still limited to the dataset used on the research and the dependence on bag of words concept instead of understanding the whole question meaning which can be done using transformer architecture.]]>
Copyrights © 2024
