Forecasting the Efficiency of Bitcoin Market Price Using Novel Apriori Algorithm and Compare the Prediction Accuracy with Linear Regression Algorithm

Abstract

In recent years, Bitcoin has gained significant attention as a leading cryptocurrency, with its price volatility drawing the interest of both investors and researchers. Predicting the future price of Bitcoin is a challenging task due to its inherent unpredictability and fluctuating market dynamics. Accurate forecasting of Bitcoin's price can provide valuable insights for traders and investors to make informed decisions. This study aims to evaluate and compare the prediction accuracy of Bitcoin prices using two distinct machine learning techniques: the Novel Apriori Algorithm and Linear Regression. This study investigates the efficiency of predicting Bitcoin prices using two machine learning techniques: the Novel Apriori Algorithm and Linear Regression. The primary objective is to forecast the Bitcoin price using these algorithms and assess their prediction accuracy. A pretest power analysis was conducted with an 80 % power level and a sample size of 20, with two distinct groups. The software implementation of both algorithms yielded an accuracy of 83.85% for the Novel Apriori Algorithm and 82.70% for the Linear Regression Algorithm. Statistical analysis, using an independent sample$t$-test, indicated a negligible difference in accuracy between the two methods ($p>0.05$), with a mean difference of$\mathbf{0. 7 6 0}$. Despite this, the Novel Apriori Algorithm demonstrated a marginally higher accuracy compared to the Linear Regression Algorithm, thus suggesting its slightly better performance in forecasting Bitcoin prices.