FORECASTING INDONESIA STOCK PRICE USING TIME SERIES ANALYSIS AND MACHINE LEARNING IN R

Abstract

This study investigated the appropriate model to predict 30 days ahead of Unilever Indonesia stock price and Telekomunikasi Indonesia stock price using time series analysis and machine learning in R, time series forecasting is a fun and interesting way to learn data science. The data is format Close Price. The goal of this project is to predict the future stock price of unilever indonesia and telekomunikasi indonesia using various predictive forecasting models and then analyze the various models. The dataset for unilever stocks is obtained from yahoo finance using Quantmod package in R. The final results that have been compared show that using the arima and neural network methods produces good accuracy values. Research and analysis of stock prices will help investors carry out investment is more accurate, investors can determine what steps will be taken, either buying a share or selling acquired shares the right step in taking an action. The data model used to predict close stock prices in this study unilever Indonesia using arima has an accuracy of 98.87%. and using neural network model has an of 98.92%. Telekomunikasi Indonesia using arima has an accuracy of 98.74%. and using neural network Model has an accuracy of 98.77% there are suggestions that can be given for further research and development. Trying to add to the existing historical data to be more complete so as to improve the accuracy of forecasting.