Rancangan Machine Learning untuk Mendeteksi Lagu Plagiat

Abstract

Plagiarism in the music industry is a serious issue that requires advanced solutions. This research proposes a Machine Learning-based system for detecting song plagiarism by combining Convolutional Neural Network (CNN) and Dynamic Time Warping (DTW). CNN is used to extract features from the visual representation of music notations, while DTW measures the temporal distance between two sequences of notations. Experimental results show that this system provides a more accurate solution with an accuracy of 92.71%, with a dataset of 4800 data points.