Predicting Market Trends with Enhanced Technical Indicator Integration and Classification Models

TL;DR

This work addresses predicting Bitcoin price direction in a volatile crypto market by framing it as a binary classification problem. It fuses historical data with a broad set of technical indicators and employs chi-squared feature selection to identify eight key features, feeding them into an XGBoost classifier. The approach achieves strong directional performance (accuracy ~92.4%, ROC AUC ~0.982) and outperforms a Logistic Regression baseline, with robust buy/sell signal generation validated on 15-minute Binance data. The study demonstrates how feature selection and gradient-boosting can enhance crypto forecasting, offering practical utility for traders while outlining avenues for expanding to more indicators and assets.

Abstract

Thanks to the high potential for profit, trading has become increasingly attractive to investors as the cryptocurrency and stock markets rapidly expand. However, because financial markets are intricate and dynamic, accurately predicting prices remains a significant challenge. The volatile nature of the cryptocurrency market makes it even harder for traders and investors to make decisions. This study presents a classification-based machine learning model to forecast the direction of the cryptocurrency market, i.e., whether prices will increase or decrease. The model is trained using historical data and important technical indicators such as the Moving Average Convergence Divergence, the Relative Strength Index, and the Bollinger Bands. We illustrate our approach with an empirical study of the closing price of Bitcoin. Several simulations, including a confusion matrix and Receiver Operating Characteristic curve, are used to assess the model's performance, and the results show a buy/sell signal accuracy of over 92\%. These findings demonstrate how machine learning models can assist investors and traders of cryptocurrencies in making wise/informed decisions in a very volatile market.

Figures (8)

• Figure 1: Research process diagram.
• Figure 2: Close price of Bitcoin from February 1, 2021, to February 1, 2022, showing trends and fluctuations over the selected timeframe.
• Figure 3: Number of Buy and Sell signals.
• Figure 4: Matrix $\mathbf{X}$ showing various technical indicators (features) selected by the $\chi^2$ statistical test. Each row represents different instances or observations, while each column corresponds to a specific technical indicator or feature. The notation $\text{RSI}_{30}^{(i)}$, $\text{MACD}^{(i)}$, $\text{MOM}_{30}^{(i)}$, etc., denote different instances/observations of the respective technical indicators.
• Figure 5: Receiver Operating Characteristic (ROC) curve comparison for XGBoost and Logistic Regression.