Implementation of Support Vector Regression for Polkadot Cryptocurrency Price Prediction

Cryptocurrency investment is an instrument that has high risk but also a greater advantage than other instruments. To make big profit, investors need to analyze cryptocurrency investments predict the price of be purchased. The highly volatile movement prices makes it difficult for those prices. Data mining process extracting large amounts information from data by collecting, using data, history relationship patterns, and relationships in sets. Support Vector Regression doing accurate predictions can overcome problem overfitting itself. Polkadot one cryptocurrencies are often used as instruments world cryptocurrencies. prediction analysis algorithm good predictive accuracy value, including daily closing namely with radial basis function (RBF) kernel cost parameters C = 1000 gamma 0.001 obtained model 90.00% MAPE 5.28 while linear kernels 10 87.68% value 6.10. It concluded through parameter tuning, formed best use 0.001. results show method quite if