Crypto currency-bitcoin Price Predictor using Linear Regression and random forest

Abstract

As a vital resource, Bitcoin has a substantial influence on economic marketplaces. This work accentuates the status of precisely estimating Bitcoin prices by seeing numerous aspects that stimulate its worth. Our goal is to recognize the most pertinent reasons and to forecast Bitcoin prices daily, showcasing outlines in Bitcoin pricing. The dataset embraces regular remarks over a year, covering the preliminary price, peak price, last price, concluding price, Bitcoin trading volume, volumes of other financial gauges, and biased values. These aspects are measured in forecasting the subsequent day's concluding price. The research employments machine learning techniques, precisely linear regression and random forest algorithms, to estimate Bitcoin prices. Exploiting an inclusive dataset from Y-Finance and Wiki, spanning five years of market data, the work accentuates the status of data preprocessing to sustain data eminence and significance. The random forest model is optimized using a grid search for hyperparameters, and performance is evaluated using Root Mean Squared Error and Pearson’s correlation coefficient. This paper highlights how these machine learning methods can improve Bitcoin price predictions, providing practical insights into model development and evaluation within volatile financial markets.