Optimization Silhouette Coefficient on Genetic Algorithm for Clustering in Auto Correction German Learning

Teaching, especially language, is very difficult; for example, speaking needs more effort. Teaching the German language need more method and tools for the student. Some tools had used in teaching language for correction, but it has a limit for the database in vocabulary. This research made a tool for helping teachers in the German language that has correct. The tools can improve the ability of teachers and students through assessment correction. The part of the application evaluates students and teachers who can speak German. This research used a genetic algorithm to classify the correction of the German language assessment. This algorithm was used to classify the correction of assessment tasks in the German language. The genetic algorithm plays a role in optimizing the initial centre of the cluster in the K-Means algorithm. The data used in this study is data from lecturers of the German language study program at Universitas Negeri Malang in 2020. The data was obtained from SIPEJAR online learning at the German Program, Universitas Negeri Malang. The results of the clustering test of students and lecture in the German language based on the correction of the assessment using the Silhouette Coefficient-GA method has a higher cluster quality of 3.24% compared to the K-Means algorithm without the genetic algorithm.