Backpropagation Network Optimization Using One Step Secant (OSS) Algorithm

Abstract

Education is one of the main indicators in national development efforts. The government in its efforts to realize national goals to educate the nation’s life has made a policy for compulsory 9-year schools, namely elementary and junior high schools. To find out how many residents use the education facilities provided by the government can be seen through school enrollment rates (SPR). A high School Participation Rate (SPR) means showing greater opportunities to access education in general. This study aims to optimize artificial neural networks with the One Step Secant (OSS) algorithm. Artificial Neural Network (ANN) is part of the artificial intelligence system (Artificial Intelligence, AI) which is one of the artificial representations of the human brain that always tries to simulate the learning process in the human brain. The sample data used for optimization is SPR Indonesia data by province. Using 4 architectural models with 5 input variables, 1 shadow layer and 1 output. The best results obtained between architectures 5-4-1, 5-8-1, 5-16-1 and 5-32-1 are architectures 5-16-1. Obtained prediction accuracy comparisons using the One Step Secant (OSS) algorithm and standard algorithms namely 96.97% and 100%. The standard algorithm is superior in accuracy, the One Step Secant (OSS) algorithm is superior in terms of iterations.