Pendekatan Algoritma Tree dalam Prediksi Populasi pada Smart Poultry

Abstract

Intelligent systems for monitoring poultry in kennels are experiencing an increasing trend in several studies. Monitoring poultry is very important in the cage so that you can find out the chickens' condition and environment in the cage. Conditions that can be monitored include the weight of the chickens, whether or not there is enough water in a day, CO2 levels in the cages, air temperature, and humidity in the cages. Several studies have been conducted studies on monitoring poultry cages using IoT-based sensors. However, people have yet to predict the poultry population for tomorrow. So this study aims to predict the number of poultry populations in kennels based on related parameters. The prediction method used in this research is a decision tree and Support Vector Machine (SVM) to see which prediction method is better. The results evaluation techniques used in this study are Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R2. The experimental results show that using the decision tree method, and the results are MSE 61987.202, RMSE 248.972, MAE 85.086, and R2 0.969. Overall the results of the decision tree method are superior to SVM.