Artificial intelligence (AI) and machine learning (ML) techniques are extensively utilized to process large data sets and identify patterns to improve the accuracy of stock market prediction. However, current models often fail to address the inherent unpredictability of financial markets, which requires innovative approaches to pinpoint critical information. This paper explores the relationship between market data and sentiment values using Association Rule Mining (ARM) to provide valuable knowledge for researchers and investors on what parameters to use to improve stock market prediction. We analyze tweets containing “bitcoin” and “investment” keywords from 2015 to 2021 using the RoBERT’a model to measure investor sentiment and its impact on market movements of major entities, including Amazon, Microsoft, and Bitcoin. In addition, we incorporate key technical indicators such as SMA, MACD, and RSI, as well as economic indicators. Our results demonstrate a moderate correlation between stock prices and significant economic indicators, particularly the Treasury yield. Furthermore, sentiment analysis reveals a slight but notable correlation between investor sentiment and stock prices. This study underscores the efficacy of ARM in financial analysis and identifies potential avenues for future research to refine predictive models for stock markets.