Pemodelan Time Series Pergerakan Harga Saham dengan Deep Learning

Abstract

The movement of stock prices is difficult to predict the direction of progress. Therefore, a method is needed that can model stock price movements in order to make it easier for investors to make investment decisions. One method that can be used is modeling time series stock price movements with deep learning. First, we will look at the Autocorrelation Function (ACF) to determine the stock data in the form of stationary or non-stationary. Furthermore, several parameters will be determined that can affect the accuracy of predicting stock price movements. A good prediction result of stock price movements is that there is no overfitting of the model and also provides predictive data results that are close to actual data. This can be seen by the small difference between MSE training and MSE testing.