| dc.contributor.advisor | Yani, Ahmad | |
| dc.contributor.advisor | Komala, Iyep | |
| dc.contributor.advisor | Arif, Chusnul | |
| dc.contributor.author | Samih, Sulthan Falah | |
| dc.date.accessioned | 2026-08-10T10:44:21Z | |
| dc.date.available | 2026-08-10T10:44:21Z | |
| dc.date.issued | 2026 | |
| dc.identifier.uri | http://repository.ipb.ac.id/handle/123456789/178042 | |
| dc.description.abstract | Produktivitas sapi perah dipengaruhi oleh kemampuan ternak mempertahankan keseimbangan fisiologis terhadap perubahan mikroklimat. Suhu udara (Ta) dan kelembapan relatif (RH) memengaruhi mekanisme termoregulasi yang tercermin melalui denyut jantung (HR) dan frekuensi napas (RR). Penilaian kondisi termal umumnya menggunakan Temperature–Humidity Index (THI), tetapi pendekatan tersebut hanya merepresentasikan kondisi lingkungan sehingga belum sepenuhnya menggambarkan respon biologis ternak. Integrasi informasi mikroklimat dan respon fisiologis diperlukan untuk menghasilkan penilaian kondisi termal yang lebih representatif.
Penelitian ini bertujuan memprediksi respon fisiologis sapi perah menggunakan Artificial Neural Network (ANN) berdasarkan Ta dan RH, mengklasifikasikan kondisi termal berdasarkan hasil prediksi fisiologis, serta mengevaluasi kesesuaiannya dengan klasifikasi berbasis THI. Penelitian dilakukan di BPPIBTSP Bunikasih, Kabupaten Cianjur, menggunakan 14 ekor sapi Friesian Holstein laktasi kedua. Data mikroklimat dan respon fisiologis disinkronkan berdasarkan waktu pengamatan untuk membangun model ANN dengan Ta dan RH sebagai variabel masukan serta HR dan RR sebagai variabel keluaran. Kinerja model dievaluasi menggunakan Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), koefisien determinasi (R²), dan tingkat kesesuaian.
Sebanyak 170 observasi berhasil dihimpun. Model ANN memprediksi HR dengan MAE 4,57, RMSE 5,85, dan R² 0,51, sedangkan prediksi RR menghasilkan MAE 4,32, RMSE 5,23, dan R² 0,69. RR menunjukkan akurasi prediksi yang lebih baik dan lebih sensitif terhadap perubahan mikroklimat dibandingkan HR. Klasifikasi kondisi termal berdasarkan hasil prediksi menunjukkan bahwa respon ternak berkembang melalui fase Respon Adaptif sebelum mencapai Heat Stress. Perbandingan dengan klasifikasi berbasis THI menghasilkan tingkat kesesuaian sebesar 85,29%, dengan ketidaksesuaian terutama terjadi antara kategori Respon Adaptif dan Heat Stress.
ANN mampu menghubungkan perubahan mikroklimat dengan respon fisiologis sehingga menghasilkan penilaian kondisi termal yang lebih mencerminkan kondisi biologis sapi perah. Pendekatan ini tidak menggantikan THI, tetapi melengkapinya sebagai dasar pengembangan sistem pemantauan kondisi termal yang lebih representatif untuk mendukung pengelolaan lingkungan kandang, peningkatan kesejahteraan ternak, dan penerapan precision livestock farming pada peternakan sapi perah. | |
| dc.description.abstract | Dairy cattle productivity depends on the ability of animals to maintain physiological homeostasis under changing microclimatic conditions. Air temperature (Ta) and relative humidity (RH) influence thermoregulation, which is reflected in changes in heart rate (HR) and respiration rate (RR). Thermal conditions are commonly evaluated using the Temperature–Humidity Index (THI); however, this approach represents environmental conditions only and does not fully reflect the biological responses of individual animals. Therefore, integrating microclimatic variables with physiological responses is necessary to provide a more representative assessment of thermal conditions.
This study aimed to predict the physiological responses of dairy cattle using an Artificial Neural Network (ANN) based on Ta and RH, classify thermal conditions according to the predicted physiological responses, and evaluate their agreement with THI-based classification. The study was conducted at BPPIBTSP Bunikasih, Cianjur, using fourteen second-lactation Friesian Holstein dairy cows. Microclimatic and physiological data collected were synchronized according to observation time to develop an ANN model with Ta and RH as input variables and HR and RR as output variables. Model performance was evaluated using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), the coefficient of determination (R²), and the Agreement Rate.
A total of 170 observations were included in the final dataset. The ANN predicted HR with an MAE of 4.57, an RMSE of 5.85, and an R² of 0.51, whereas RR prediction achieved an MAE of 4.32, an RMSE of 5.23, and an R² of 0.69. RR showed higher prediction accuracy and greater sensitivity to microclimatic variation than HR. Physiological classification demonstrated that thermal responses developed progressively through an Adaptive Response stage before reaching Heat Stress. Comparison with THI-based classification produced an Agreement Rate of 85.29%, with most disagreements occurring between the Adaptive Response and Heat Stress categories.
The ANN successfully linked microclimatic variation with physiological responses, providing a thermal assessment that more closely reflected the biological condition of dairy cattle. This approach does not replace THI but complements it, providing a basis for developing more representative thermal monitoring systems to support barn environmental management, improve dairy cattle welfare, and advance precision livestock farming. | |
| dc.description.sponsorship | | |
| dc.language.iso | id | |
| dc.publisher | IPB University | id |
| dc.title | Prediksi dan Klasifikasi Respon Fisiologis Sapi Perah Laktasi Menggunakan Artificial Neural Network pada Dataran Tinggi di BPPIBTSP Bunikasih, Cianjur | id |
| dc.title.alternative | Prediction and Classification of Physiological Responses in Lactating Dairy Cattle Using Artificial Neural Network in the Highland Area at BPPIBTSP Bunikasih, Cianjur | |
| dc.type | Tesis | |
| dc.subject.keyword | Artificial Neural Network | id |
| dc.subject.keyword | Kondisi Termal | id |
| dc.subject.keyword | mikroklimat | id |
| dc.subject.keyword | respon fisiologis | id |
| dc.subject.keyword | Sapi Perah | id |
| dc.subtype | Theses | |