Optimization of Random Forest with Genetic Algorithm for Determination of Assessment

The Government of the Republic of Indonesia has issued a self-study regulation in which there are many parameters to determine the graduation status of MBMK participating students. Choosing the graduation of MBKM learning is not only based on rules from managers, universities, faculties, departments, or students. The most fundamental problem is that there are no standard rules regarding graduation status and graduation parameters in MBKM learning. Based on these problems, this study applies the random forest method optimized with genetic algorithms to classify students' graduation status accurately and quickly. After testing, the average accuracy with the random forest method optimized by the genetic algorithm is 88.58%, the average precision is 77.34%, the recall average is 57.68%, and the f1-score is 66.54
^%
, with the parameter size population of 21, 3 iterations, crossover rate value 0.8, mutation rate value 0.4, and several features 4. The evaluation results of accuracy, precision, recall, and f1-score obtained in this study show that the genetic algorithm can find optimal parameters. The random forest result in higher accuracy, precision, recall, and f1 scores