PREDIKSI BIAYA PRODUKSI TERNAK BABI MENGGUNAKAN ALGORITMA KNN PADA PETERNAKAN TOLU WEI

Abstract

Indonesia, as an agrarian country, places the livestock sector as an essential component in supporting economic development and food security. In the East Nusa Tenggara (NTT) region, particularly in East Sumba Regency, pig farming holds high economic, social, and cultural significance. Peternakan Tolu Wei is one of the growing pig farms in the area, raising hundreds of pigs. However, inefficient production cost planning often leads to unnecessary spending. This study aims to predict pig farming production costs using the K-Nearest Neighbors (KNN) algorithm by utilizing historical data from the past three years. The analyzed variables include feed costs, labor wages, healthcare, electricity/water, and housing costs. The analysis process is carried out using RapidMiner software. The evaluation results of the model show a high level of accuracy, with RMSE = 0.370, Absolute Error = 0.282, and Relative Error = 1.31%, indicating that the model is highly accurate. The predicted total monthly operational costs for 2025 range between IDR 19.9 million and IDR 21.4 million, reflecting a consistent spending pattern. This model is not only beneficial for Peternakan Tolu Wei but also has the potential to be applied to other pig farms with similar characteristics, in order to support more efficient cost management in livestock production.

ABSTRAK