PERBANDINGAN PERFORMA MODEL MACHINE LEARNING SUPPORT VECTOR MACHINE, NEURAL NETWORK, DAN K-NEAREST NEIGHBORS DALAM PREDIKSI HARGA SAHAM

Sudriyanto, Sudriyanto
Syahro, Fatimatus
Fitriani, Novi

JAR'S (Journal of Advanced Research in Informatics)
Vol 2 No 1 (2023): JARS : Journal of Advanced Research in Informatics

This study aims to analyze the performance of three prediction models, namely K-Nearest Neighbors (K-NN), Neural Network (NN), and Support Vector Machine (SVM), in predicting the stock price of PT Astra International Tbk (ASII.JK). The research encompasses the initial stages through evaluation using optimal parameters for these three algorithms. The research findings reveal that the K-NN prediction model has the lowest Root Mean Square Error (RMSE) value, with a value of 0.037, indicating the most accurate prediction compared to the other models. Despite the NN model having an RMSE of 0.048, which is higher than K-NN, it still provides reasonably accurate predictions. Meanwhile, the SVM model has an RMSE of 0.075, indicating a higher level of error in its predictions. Based on these results, the recommendation is to utilize the K-NN model as the preferred choice for predicting the ASII.JK stock price.