SENTIMENT ANALYSIS ARTICLE NEWS COORDINATOR MINISTER OF MARITIME AFFAIRS USING ALGORITHM NAIVE BAYES AND SUPPORT VECTOR MACHINE WITH PARTICLE SWARM OPTIMIZATION

News has become a basic human need along with the development of technology and the internet. The

development of technology and the internet is causing the change of publication pages from a print media

to the internet. The use of online media today is not only for reading news articles, but also can be used to

see the issues that occur can even be used to see the performance of a political figure. The classification of

the contents of news articles into a new knowledge that is a negative or positive conclusions about the

content of the news contained in a news site. It is possible by using sentiment analysis that is by document

classification with text mining. The algorithm used in this research is Naive Bayes and Support Vector

Machine with Particle Swarm Optimization. NB has an accuracy value of 89.50% with AUC of 0.500

while the NB PSO obtains an accuracy of 92.00% with AUC of 0.550. SVM has an accuracy value of

87.50% with AUC of 0.979, while SVM PSO has an accuracy value of 90.50% and AUC of 0.975. The

best application of optimization in this model is NB PSO can provide solution to the classification problem

in this case of sentiment analysis. NB PSO algorithm provides solutions to the analysis of

sentiments from the content of various online media news optimally.

Unduhan