<![CDATA[YouTube users can create, watch, and share videos for free. Interaction between viewers occurs through the comment feature, which can be positive or negative. The frequent appearance of negative comments on the youtube application on the google play store can have an effect on these accounts. But to find out how much negative comments on the account are needed, an SVM algorithm is needed. This study aims to determine the sentiment towards the youtube application on the google play store using Machine Learning with the SVM algorithm. The data taken is 4996 review data which is then preprocessed so that the remaining data becomes 4993 data that can be processed. Data labelling is done automatically based on the review rating score. The results of data labelling are divided into 3 classes, namely positive classes as many as 1083, negative classes as many as 3365 and neutral as many as 545. Classification and evaluation are carried out using the SVM method. Based on the training and testing data comparison value of 9: 1, the results obtained an accuracy rate of 75% then negative class precision of 76% and negative class recall of 97% and K-Fold Cross Validation testing using a value of K = 10 with an average accuracy of 0.75 or 75%.]]>
Copyrights © 2024
