NEWS TOPIC CLASSIFICATION WITH MACHINE LEARNING USING SEMI-SUPERVISED LEARNING

Nowadays, the amount of news has increased every day. They also grow rapidly. This sit-uation makes it difficult for the editor to categorize manually and this manual categorization makes the categories incorrect for the news. They also wasted the time choosing the category for the news which were published. For that problem, we proposed news categories classification using machine learning with semi-supervised learning with pseudo labeling method. We also proposed an ensemble learning algorithm called Adaboost that can boost the data from misclassified. Hopefully, this algorithm and method will solve the problem effectively. For the result of this research, Semi-supervised learning also can improve performance. By the data from supervised learning, we can see the precision, recall, and F1-Score with the same ratio in semi-supervised learning and the performance improved well. If the data varied with the same learning ( supervised or semi-supervised learning ) the performance of learning also in-creased. Although there is the result that is not increased, it is just 1-5% only a small difference between the