Parent Coin based Cryptocurrency Price Prediction using Regression Techniques

Abstract

Cryptocurrencies are becoming increasingly popular day by day among people. It is providing more features than traditional banking or money can provide, hence attracting investors all around the world. But people are still hesitant to use this technology because of the high price volatility are variable markets. The majority of the forecasting solutions reported have some level of error and cannot predict the price accurately due to randomness. The proposed model consists of using parent coins as a parameter for price prediction to overcome market volatility. Moving average is used as a data preprocessing technique for the effective prediction of Litecoin and Dogecoin prices. Also, in this, we are using Decision Tree, Random Forest, Extra- Tree-Regressor and Ridge regressor models for predicting the close price of cryptocurrencies. From the result, we can see that our proposed model performs better than predicting without using the parent coin feature.