<![CDATA[This study aims to classify student achievement levels based on academic ability in mastering subject matter using the K-Nearest Neighbor (K-NN) method. This study is motivated by the limitations of the student grade data processing system which is still done manually using Microsoft Excel, where the process of adding and grouping grades into low to high categories takes a long time and makes it difficult for teachers to identify student achievement levels, such as Good, Sufficient, and Poor categories. The data used consists of 134 students with 11 subject attributes as input variables in the classification process. The results show that from 134 student data, 90 students are classified into the Good category, 20 students are classified into the Sufficient category, and 24 students are classified into the Poor category. Testing using RapidMiner shows that the K-NN method obtains an accuracy level of 89.55%, which indicates that this method is effective in grouping student achievement levels. Performance evaluation is carried out using a Confusion Matrix to compare the classification results with actual data. Testing was conducted 10 times with a total of 134 student data obtained, the K-Nearest Neighbor (KNN) algorithm produced an accuracy of 89.55% and a kappa value of 0.763. These figures indicate that the model has excellent classification performance and a high level of reliability. The resulting model was then developed into a web-based prediction system that was tested using the expert system method through the Black-box Testing approach and obtained a feasibility level of 83.44%.]]>
