CodelessML: A Beginner's Web Application for Getting Started with Machine Learning

Authors : Hanif Noer Rofiq, Galuh Mafela Mutiara Sujak

Pages : 582-599

Doi:10.18009/jcer.1506864

View : 130 | Download : 106

Publication Date : 2024-10-21

Article Type : Research Paper

Abstract :Building machine learning models requires intensive coding and installation of certain software. This is frequently a barrier for beginners learning about machine learning. To overcome this situation, we present CodelessML, a reproducible web-based application designed for Machine Learning beginners due to its coding-free and installation-free design, published under Code Ocean capsule. It provides a common workflow that eases the process of building Machine Learning models and using the model for predictions. Using the Agile method, CodelessML was successfully built using Python, Anaconda, and Streamlit It. By using CodelessML, users can get a walkthrough and interactive experience of building machine learning through a simplified machine learning process: exploratory data analytics (EDA), modelling, and prediction. The impact of the software was evaluated based on feedback from 79 respondents, which showed that based on a 5-scale Likert, CodelessML received average ratings of 4.4 in accessibility, 4.3 in content, and 4.4 in functionality. CodelessML serves as an accessible entry point for learning machine learning, offering online, free, and reproducible features.
Keywords : Machine learning, learning, barrier, software

ORIGINAL ARTICLE URL