Implementasi Web Klasifikasi Suasana Hati Berdasarkan Potongan Lagu dengan Memanfaatkan Convolutional Neural Network

Abstract

Music is often used to accompany the user according to his heart condition. So users often create a playlist by adjusting the mood they are feeling. However, there are some users who have had difficulties in making playlists because in making a playlist it has to be done manually, that is, listening to music one at a time, wasting a lot of time. Therefore, the author conducted research on the classification of the mood contained in music and created a system that works to help classify music automatically by using one method that is part of deep learning, the method mentioned by the author is the Convolutional Neural Network (CNN) method. As for the data used by the investigator in this study is music data with a lot of data amounting to 400 data, on such data is done preprocessing data by cutting the duration of music and converting music into image. The next step is to split the data, dividing it into training data and test data. The training data is divided by 80% and the test data is also split by 20% of the total datasets used by the author. The results of this data division were used to build a model using the CNN model. The accuracy results obtained in this study were 95% for the training accurately and 68% for the data validation accurate.