Universitas Dian Nuswantoro conducted a journal activity on nitrous oxide (N2O) emissions from agricultural activities. The activity involved Muslikh, Ahmad Rofiqul; Andono, Pulung Nurtantio; Marjuni, Aris; and Santoso, Heru Agus as participants. The activity relates to Sustainable Development Goals 2 and 13. It proposed an ensemble framework combining Isolation Forest, DBSCAN, and One-Class SVM to enhance outlier detection in N2O emission datasets.
The dataset consisted of 2,246 rows and 21 columns. It was preprocessed to address missing values and normalize the data. Outlier detection was performed using each method individually and was followed by integration through hard and soft voting. Isolation Forest identified 113 outliers, DBSCAN detected 1,801 outliers, and One-Class SVM found 118 outliers. Hard voting identified 165 outliers, while soft voting detected 734 outliers. The process resulted in a refined dataset for subsequent modeling.
The ensemble approach improved the accuracy of the XGBoost model for N2O emission prediction. The best results were obtained using Random Search Cross Validation for hyperparameter tuning with a test size of 20%. The reported results were a CV MSE of 0.0215, an MSE of 0.0144, an RMSE of 0.1200, an MAE of 0.0723, and an R2 of 0.6750.
Through this activity, UDINUS addressed the combination of multiple outlier detection methods to enhance data quality and model performance, supporting more reliable predictions of N2O emissions. The source does not provide additional information about the activity beyond the methods and results described.