Klasifikasi Retensi Tenaga Kerja Panen Kelapa Sawit di Perusahaan XYZ Menggunakan Algoritma Random Forest

The palm oil industry plays an important role in Indonesia's economy, making retention — the company's ability to retain its workforce, particularly palm harvest workers — a key factor for the stability and performance of Company XYZ. This study employs the Random Forest algorithm to classify and identify the main factors influencing the retention of palm harvest workers. The dataset, sourced from the HR department, consists of 370 workers' data collected over one month, comprising both numerical and categorical features. The preprocessing stages include data cleaning, feature selection and creation, data transformation, and the application of SMOTE to address class imbalance. The data is split into a 70:30 ratio for training and testing. The model was tested under two scenarios: using default parameters, achieving an accuracy of 98%, and with optimization through GridSearchCV, boosting accuracy to 99%. Feature analysis revealed that the key factors affecting retention include Age, Monthly Harvest, Weekly Harvest, Daily Harvest, Total Salary, Actual Productivity and other contributing factors, with a consistent pattern of importance. The results demonstrate that the Random Forest algorithm effectively classifies the retention of palm harvest workers. These findings are expected to provide data-driven insights that HRD can act upon to classify the retention of palm oil harvest workers.