The development of business technology has made people interested in following online-based businesses, one of which is stock trading. In conducting stock trading, a thorough analysis of market conditions is required to avoid losses. However, not everyone can analyze the stock market, machine learning plays an important role in helping this analysis. Researchers used a simulation method with historical data on PT. Telkom Indonesia’s shares for 2018-2022 to compare the algorithm's performance. In the experimental stage, modeling and simulation were carried out using RapidMiner Studio, with 20 different scenarios, including variations in the split ratio of training and test data. The simulation results were analyzed using the RMSE evaluation metric, which measures the difference between predicted and actual values. The experimentation results show that the Naïve Bayes algorithm has the best performance with the lowest RMSE value, which is 0.839, at a split ratio of 0.7 to 0.3. On the other hand, the Support Vector Machine algorithm showed the worst performance with the highest RMSE, which is 91.043, at the same split ratio.
Copyrights © 2024