Jerami padi merupakan sumber glukosa yang boleh digunakan sebagai bahan mentah untuk penghasilan pelbagai produk yang bernilai tinggi. Hidrolisis enzim selulosa daripada jerami padi adalah proses yang kompleks kerana ia melibatkan perencatan dan penyahaktifan enzim ketika proses hidrolisis. Artificial Neural Network (ANN) ialah kaedah yang berkesan dalam menghasilkan model jangkaan untuk proses yang melibatkan kinetic reaksi yang kompleks yang mana sukar untuk dimodelkan mengunakan kaedah yang lebih tradisional. Kajian ini dijalankan untuk mengkaji aplikasi Artificial Neural Network (ANN) sebagai kaedah untuk meramalkan penghasilan glukosa melalui hidrolisi enzim daripada jerami padi dan membandingkan antara data yang diperolehi dari ANN dengan kaedah Tindak Balas Permukaan (RSM) dan data yang diperolehi daripada eksperimen. Neural network yang mempunyai satu hidden layer telah dilatih untuk meramalkan penghasilan glukosa. Nilai R2, MSE dan ARD telah diperolehi dari model neural network. Daripada perbandingan yang dilakukan, boleh dikonklusikan bahawa ANN menghasilkan ramalan yang lebih tepat berbanding RSM. Keputusan yang diperolehi menunjukkan bahawa ANN merupakan kaedah yang sesuai bagi mereka bentuk hidrolisis enzim.
_______________________________________________________________________________________________________
Rice straw is a source of glucose which can be used as a raw material for production of many high value products. Enzymatic hydrolysis of cellulose from rice straw is a complex process because of a number of inhibition and enzyme inactivation reactions which happen during hydrolysis. Artificial Neural Networks (ANNs) are very effective in developing predictive models for processes involving complex reaction kinetics that would otherwise be difficult to be modeled by more traditional deterministic approaches. The present investigation was carried out to study the application of Artificial Neural Network as a tool for predicting glucose production by enzymatic hydrolysis of rice straw and comparison of the predicted data with Response Surface Methodology (RSM) model and experimental values. A feed forward neural network with one hidden layer was trained and used to predict the glucose production. The performances of the model were evaluated using the coefficient of determination, mean square error and average relative deviation. The predictive model show a good result as the coefficient of determination, 0.8361 was obtained with small value of mean square error, 0.1947 and 5.644 of the average relative deviation. It is clearly shows that ANN successfully predicts and gave a good prediction on the enzymatic hydrolysis for the production of glucose. The results from predictive model were then compared with Response Surface Methodology (RSM) model. From the comparison, it was concluded that the neural network has a better prediction accuracy than the RSM. The obtained results show that the ANN can be a useful method for the design of the enzymatic hydrolysis.