Data Analysis of Bitcoin Price Trends Using KNN Prediction Models

Abstract

This study investigates Bitcoin price trends and evaluates the effectiveness of the K-Nearest Neighbors (KNN) algorithm for predicting price movements in the cryptocurrency market. Leveraging a decade of historical Bitcoin price data, trading volume, and market capitalization, the research assesses the accuracy and reliability of KNN in capturing the complex and volatile nature of Bitcoin price dynamics. The methodology includes data preprocessing, exploratory analysis, and predictive modeling with hyperparameter optimization. The findings reveal that while KNN achieves moderate accuracy (53%), it performs better in identifying price decreases (Class 0) with a recall of 66% compared to price increases (Class 1) with a recall of 40%. The study also highlights key challenges, including Bitcoin's high volatility and multicollinearity among features like Moving Averages. To improve prediction accuracy, the research recommends feature expansion, advanced modeling techniques (e.g., LSTM networks), and the integration of external factors such as market sentiment and macroeconomic indicators. These results contribute to the growing body of knowledge in cryptocurrency forecasting, providing insights for investors, traders, and researchers to navigate the complex cryptocurrency landscape.