Data Mining Approach to Predict Air Pollution in Makassar

Abstract

Air pollution level in Makassar has increased based on data from 2018 to 2019. There were 646 data obtained from the Ministry of Environment and Forestry data archive through the official site, there as five variables in data training, particulate matter (PM10), sulfur dioxide (SO2), carbon monoxide (CO), nitrogen dioxide (NO2), and ozone (O3). Lack of information on air pollution causes the people become unaware on their personal health. There is an effective analysis method for exploring data. This research used knowledge discovery technique in databases in data mining to facilitate decision making. Finally, continuing from the results of previous studies where the prediction of air pollution levels used Naïve Bayes algorithm, this research predicts the level of air pollution using the K-Nearest Neighbor Algorithm to classification data testing and data training with an accuracy rate of 96%, a precision value of 97% and also a recall value of 100%.