This chapter presents a comprehensive analysis of brain cancer gene expression datasets through binary and multi-class classification using the CatBoostClassifier, enhanced by Principal Component Analysis (PCA) for dimensionality reduction. The result and discussion section elucidates key findings, trends, and the efficacy of the methodologies employed. Utilizing Volcano Plots, we identified significant biomarkers that differentiate gene expression between cancerous and normal tissues, facilitating the discovery of potential diagnostic targets. In the multi-class classification, the model effectively distinguished between various brain cancer types, including ependymoma, glioblastoma, medulloblastoma, and pilocytic astrocytoma, achieving an overall accuracy of 87%. Conversely, the binary classification model exhibited remarkable performance, attaining 100% accuracy, precision, recall, and F1-score in distinguishing tumors from normal samples. This chapter underscores the potential of machine learning techniques in advancing brain cancer diagnostics and improving patient outcomes.