The increasing complexity and dimensionality of modern datasets have necessitated the use of advanced machine learning techniques for effective data analysis. Our result highlights the accuracy comparison between supervised and unsupervised models, showing that supervised techniques such as Support Vector Machine and Random Forest achieve higher accuracy values ranging from 0.85 to 0.91, while unsupervised models such as K-Means and DBSCAN demonstrate lower performance between 0.70 and 0.78. Further our study illustrates the relationship between error rate and data complexity, revealing that supervised models maintain lower error at lower complexity levels but become more sensitive as complexity increases, whereas unsupervised models show more stable but higher error rates. Also, our data provides a visualization of clustering patterns, demonstrating the capability of unsupervised learning techniques to identify underlying structures in data without labeled information. However, the lack of clear cluster separation highlights challenges in handling highly complex datasets. Also, our report presents the computational efficiency of different models, showing that supervised models require higher processing time, while unsupervised techniques offer faster performance but lower accuracy. Overall, the findings indicate that supervised learning techniques are more suitable for high-accuracy prediction tasks when labeled data is available, whereas unsupervised methods are valuable for exploratory data analysis and pattern recognition. The study emphasizes the importance of selecting appropriate techniques based on dataset characteristics and application requirements and highlights the potential of hybrid approaches for improving overall performance.