Learning is considered to be the best solution for teaching and learning activities in the midst of the covid-19 pandemic. Online learning can be done by providing material to students in the form of videos and powerpoint through Google Meet, Google Classroom, Zoom and others. The importance of student satisfaction is one of the benchmarks for success and quality of learning in the network. Quality learning will have a high level of satisfaction for its users. Therefore, it is necessary to classify student satisfaction with online learning to determine the quality of education. The indicators used to measure the level of student satisfaction are the relevance of learning, the attractiveness of learning in the network, effectiveness, efficiency and productivity of learning. Random Forest was chosen because the process is learning and classification very simple and fast and generally has a high level of accuracy. Random Forest able to classify data that has incomplete attributes, and able to process quite a lot of data. In this study, an accuracy test will be carried out to determine how accurate the method is random forest in classifying student satisfaction with online learning. The results of the accuracy test random forest obtained an accuracy of 81%, precision '0' 0.83, precision '1' 0.81, recall '0' 0.79, recall '1' 0.84, MAE 0.192, MSE 0.192, and RMSE 0.438.

Keywords: classification, online learning, random forest