Prediksi Harga Saham Harian Menggunakan Cascade Forward Neural Network (CFNN) Dengan Particle Swarm Optimization (PSO)

Abstract

Artificial Neural Network (ANN) is an information-processing system which characteristics and how it worked are similar to human brain. ANN can be used as one of the forecasting method without having to fulfill various assumption suc as ARIMA method. Feed Forward Neural Network (FFNN) is one of the ANN model that have a simple network architecture, where the connection between input and output occurs indirectly through the hidden layer. Cascade Forward Neural Network (CFNN) is ANN that its architecture similar to Feed Forward Neural Network (FFNN), but there is also a direct connection between input layer and output layer. CFNN can be combined with various optimization method such as Particle Swarm Optimization (PSO). Particle Swarm Optimization (PSO) inspired by flocks of birds flying in groups. Combination between Cascade Forward Neural Network (CFNN) and Particle Swarm Optimization (PSO) aimed to obtain optimum weight that minimize error. In this study, the data used was daily stock price of PT. XL Axiata Tbk. Based on the calculation results using CFNN with PSO obtained MAPE of training data equal to 1,8248% and MAPE of testing data equal to 2,3136%, which MAPE value is still less than 10% so the accuracy of the model is considered very good.