Improving the Efficiency of Bitcoin Market Price Prediction Using Novel Decision Tree Algorithm and Compare the Prediction Accuracy with K-Nearest Neighbor Algorithm(KNN)

Abstract

Bitcoin, as one of the leading cryptocurrencies, has garnered significant attention due to its highly volatile market behavior. Accurate prediction of Bitcoin prices is crucial for investors, traders, and financial analysts who seek to navigate the uncertainties of this digital currency market. In recent years, machine learning algorithms have emerged as powerful tools for forecasting financial trends, with various models being tested for their ability to predict Bitcoin prices. Among these, the Novel Decision Tree Algorithm and K-Nearest Neighbor (K-NN) algorithm have been recognized for their potential in making accurate predictions. This study aims to improve the efficiency of Bitcoin market price prediction using the Novel Decision Tree Algorithm and to evaluate its performance in comparison with the K-Nearest Neighbor (K-NN) Algorithm. The analysis was conducted with a sample size of 20 for both groups, and a pretest power analysis was performed at an 80% power level. The software implementation of both algorithms resulted in a prediction precision of 87.80% for the Novel Decision Tree Algorithm and 86.91% for the K-NN Algorithm. To assess the statistical significance of the results, an independent sample t-test was conducted, revealing that the difference in accuracy between the two algorithms was statistically negligible, with a value of 0.745 ($p > 0.05$). Despite the small difference, the Novel Decision Tree Algorithm outperformed the K-NN Algorithm in terms of accuracy, demonstrating a higher precision in Bitcoin price prediction.