This papers delves into the critical parameters of decision tree-based ensemble models, such as the number of trees (T), which necessitates careful user selection. We investigate the impact of T on model performance, aiming to determine whether setting it to the largest computationally manageable value is optimal or if a smaller T might yield better results with proper tuning. Furthermore, we extend traditional decision tree classifiers to accommodate uncertain data, a common occurrence in real-world scenarios due to factors like measurement errors, data staleness, and multiple measurements. Our research addresses the uncertainty problem associated with decision tree-based methods, providing valuable insights into its consequences and implications. The research context revolves around stock market prediction, a challenging and dynamic domain crucial for financial decision-making. Stock market behavior is influenced by various macroeconomic factors, making precise predictions complex. This study categorizes stock price prediction into trend classification and price forecasting and emphasizes the significance of incorporating fundamental analysis methods to forecast stock prices accurately. Additionally, the role of news in stock market prediction is explored, highlighting the challenges posed by unstructured news data. In summary, this work contributes to the understanding of ensemble methods, decision tree-based algorithms, and their application in stock market prediction, while also addressing the novel challenge of uncertainty in predictive modeling.