Forecasting onion armyworm using tree-based machine learning models

In the Philippines, province of Nueva Ecija produces fifty-four percent its annual onion production. However, level growth production was reduced; since outbreak 2016, armyworms destroyed thousands hectares farms resulting in a loss billions pesos, which lead to decline harvest. this study, we develop machine learning models forecast an help evade or reduce damage caused by armyworm outbreak. Climatic data; particularly Maximum temperature, Minimum Temperature, Ultraviolet Index, Humidity, Cloudiness, Wind Speed, Sun Hours, Rainfall, and Pressure from Philippine Atmospheric, Geophysical Astronomical Services Administration (PAGASA) occurrences data Provincial Agriculture Office (PAO) used as dataset for study Using Tree-based Decision Tree Random Forest. Binary classifiers were developed evaluated occurrence non-occurrence use feature importance distinguish most critical climatic features that significantly contribute forecasting Ecija. These tree-based produced satisfactory results, with Forest model exhibiting better capability than model.