<![CDATA[Abstrak. Nilai Tukar Petani (NTP) merupakan alat untuk mengukur kemampuan tukar produk yang dijual petani dengan produk yang dibutuhkan petani dalam produksi dan konsumsi rumah tangga. NTP khususnya pada subsektor Tanaman Pangan, seringkali berfluktuasi setiap bulannya, sehingga dipandang perlu untuk diprediksi dengan tepat, agar dapat membantu pemerintah dan pihak terkait mempersiapkan tindakan-tindakan pencegahan seperti menjaga kestabilan harga produksi pertanian dan mengendalikan harga-harga biaya usaha pertanian. Paper ini menguji penerapan Jaringan Syaraf Tiruan (JST) Backpropagation untuk memprediksi NPT pada sub sektor Tanaman Pangan di Provinsi Kalimantan Selatan. Pengujian dilakukan menggunakan 84 data latih dan 36 data uji. Data masukan berupa data deret waktu NPT subsektor Pangan Provinsi Kalimantan Selatan selama 12 bulan sebelumnya untuk memprediksi NPT setiap bulannya selama periode 12 bulan mendatang. Hasil uji menunjukkan nilai presentase error (MAPE) 0,97, atau diperoleh persentase akurasi prediksi sebesar 99,03 %.Kata Kunci: Prediksi; Nilai Tukar Petani; Subsektor Tanaman Pangan; Jaringan Syaraf tiruan; Backpropagation Abstract. Farmer's Exchange Rate is a tool to measure the ability to exchange products sold by farmers with products needed by farmers in household production and consumption. Farmer's Exchange Rate, especially in the Food Crops sub-sector, often fluctuates every month, so it is deemed necessary to predict accurately, in order to assist the government and related parties in preparing preventive measures such as maintaining stability in agricultural production prices and controlling agricultural costs. This paper examines the application of Backpropagation Artificial Neural Networks to predict Farmer's Exchange Rate in the Food Crops sub-sector in South Kalimantan Province. The test was carried out using 84 training data and 36 test data. The input data is in the form of time series data for the Farmer's Exchange Rate of the Food sub-sector of South Kalimantan Province for the previous 12 months to predict the Farmer's Exchange Rate every month for the next 12 month period. The test results show the percentage error value (MAPE) is 0.97, or the percentage of prediction accuracy is 99.03%.Keywords: Prediction; Farmer's Exchange Rate; Food Crops Subsector; Artificial Neural Networks; Backpropagation]]>
