Prediksi Kualitas Udara Menggunakan Algoritma K-Nearest Neighbor

Abstract

3 sampai K = 9, didapatkan bahwa nilai K = 7 mempunyai performa terbaik dengan akurasi tertinggi sebanyak 96%, presisi 92%, recall 95%, dan f-measure 93%. Kata Kunci: data mining, knn, kualitas udara, prediksi ABSTRACT In everyday life, air is used for breathing by living things. Clean air contains many benefits for life. However, in reality the air in nature is not always clean so that it can cause a decrease in air quality. Air quality like this can have an impact on human health and the surrounding environment. Air quality in DKI Jakarta can be known through the Air Pollutant Standard Index (ISPU). This study aims to predict the air quality in DKI Jakarta based on ISPU data. Prediction is done using data mining techniques with the classification method Algorithm that functions in making predictions, namely K-Nearest Neighbor (KNN), where this algorithm is an algorithm that classifies new object classes based on their closest neighbors. The data used in the study amounted to 450 data, then the data was divided into 2, test data and lattice data. This study also evaluates the algorithm model which includes the values of accuracy, precision, recall and f-measure for each tested K value. This measurement aims to determine the optimal parameters in the dataset used. As for the data obtained from testing the values of K = 3 to K = 9, it was found that the value of K7 has the best performance with the highest accuracy as much as 96%, precision 92%, recall 95%, and f-measure 93%.