<![CDATA[Online news is a report that discusses an event packaged by the media as a means of publication in the form of news that can be accessed online. The 2019 election was one of the topics that was very much discussed at that time. In its implementation, the 2019 election reaped many critical notes related to the implementation and the issue of the integrity of the election itself. In this study, researchers took titles from various online news portals related to the 2019 election for sentiment analysis. The classification process is divided into three classes, namely positive, neutral and negative. The data used in this study amounted to 395 records. The stages carried out in this study are preprocessing which includes casefolding, remove punctuation, handling whitespace, stopword removal, stemming and tekonization. Next is handling imbalanced data to balance the number of classes. After going through the preprocessing stage, the next is the processing stage using the logistic regression method and randomized search cross validations. This is used to find the best parameters where the results of these parameters are then carried out by fitting the model. The results of the combined logistic regression method and randomized search cross validation show an accuracy score of 86%.]]>
Copyrights © 2023
