PENERAPAN ALGORITMA PEMBELAJARAN PERCEPTRON UNTUK PREDIKSI SUHU EFEKTIF SASARAN DALAM KANDANG AYAM BROILER TERTUTUP

Abstract

Broiler chickens are broiler breeds that have a relatively fast growth of about 4-5 weeks. The growth of broiler chickens is influenced by several aspects, one of which is cage management. In cage management there is a target effective temperature which is one of factors that support the growth of broiler chickens. The target effective temperature for growing comes from measured temperature combined with air humidity and measured wind speed. In this article, we will discuss how the perceptron algorithm can predict the target effective temperature in the broiler chicken closed house cage systems. The goal is to find a network model that can predict target effective temperatures with high accuracy based on variable measurements. Furthermore, the result of network model will be used to regulate the conditions of the coop according to the needs of the comfort of chickens. Based on the results of this study, the perceptron algorithm with a single layer provides a good network model of target effective temperature to regulate cage conditions.