Polkadot Cryptocurrency Close Price Prediction Using Machine Learning

Cryptocurrency investment can be highly profitable with high risk. Buyers need to be able to predict the price to know when to buy at the lowest price and to sell at the highest price. The goal of the research was to conduct a study to apply various Machine Learning (ML) models such as Random Forest Regression (RFR), K-Nearest Neighbour Regression (KNNR), Decision Tree Regression (DTR), and XGBoost to predict polkadot cryptocurrency close price. Research methods were conducted by data collection, preprocessing, normalization using Min-Max Normalization, utilizing ML models that have been proposed, and evaluated. Based on the results obtained from this research, all models have good evaluation metrics scores in terms of Mean Absolute Percentage Error (MAPE), R2 score, dan Mean Squared Error (MSE). The best model is achieved by RFR with a MAPE value of 0.0338 compared to the SVR-RBF model with a MAPE value of 0.0528.