Sikana, Arina Mana (2026) Analisa Klasifikasi Nilai Pada Mata Pelajaran Sekolah Menengah Atas Negeri 1 Enok Menggunakan Metode K-Nearest Neighbor (K-Nn). Other thesis, Universitas Islam Riau.

Text skripsi_203510210_watermark.pdf - Published Version Restricted to Registered users only Download (7MB) | Request a copy

Abstract

Education plays a crucial role in developing the quality of human resources. However, the management of student grade data at SMA Negeri 1 Enok is currently conducted manually, increasing the risk of input errors and inefficiency in performance analysis. This research aims to investigate the effectiveness of the KNearest Neighbor (K-NN) algorithm and compare it with a Rule-Based System (RBS) for student achievement classification. The primary focus of this study is to empirically prove the limitations of machine learning when applied to problems with fixed, deterministic rules. The testing was conducted using 113 grade data samples with an 80:20 training and testing data split. The results indicate that the Rule-Based System (IF-ELSE logic) achieved absolute accuracy of 100%, while the K-NN algorithm only achieved 86.9%. The lower accuracy of K-NN compared to RBS is due to the probabilistic nature of the K-NN algorithm, which operates based on proximity between neighbors, thereby introducing a risk of misclassification on data that requires rigid value thresholds. This research concludes that using complex algorithms like K-NN is inappropriate and provides no added value compared to simple logical rules for student grade classification.

Actions (login required)

View Item