Komparasi Algoritma Optimization Particle Swarm Dan Support Vector Machine Pada Neural Network Untuk Prediksi Harga Saham
Andri Pramuntadi, Ari Budi Riyanto, Imam Adi Nata, Maya Mars 23 Januari 2018 Teknik Informatika

Abstract / Intisari :
This research is started from previous research about stock prediction using artificial neural network which is optimized using Particel Swarm Optimization (PSO). In this study the results of the study will be comparable with Alogaritma Support Vector Mechine which is optimized by using Particel Swarm Optimization (PSO). This comparison uses the dataset divided into 2, which is divided into 2 periods ie for short duration and long term. The first experiment was carried out by training the dataset with NN and SVM with the two data, then the data was re-examined by selecting feature selection with each algorithm ie PSO-NN and PSO-SVM. The dataset resulting from the training is then re-tested for further training. From the experimental results obtained for short-term datasets with PSO-NN obtaining a rmse of 0.467and rmse 0,455for PSO-SVM, while for long term rmse obtained from PSO-NN of 0.364 and PSO-SVM rmse obtained at 0.658. From these comparisons it is concluded that on the few datasets the PSO-SVM is able to obtain better accuracy while for more PSO-NN datasets are more explicit in predictions.

File PDF :
   Andri Pramuntadi, Ari Budi Riyanto, Imam Adi Nata, Maya Marselia-64-stmikelrahma.pdf

Tag Jurnal :