This study presents a robust method for uncovering hidden patterns within data. Initially, the preprocessing phase eliminates irrelevant data to reduce computation time. Following this, feature selection is conducted using the Principal Component Analysis (PCA) method to identify the most pertinent features. Once optimal features are selected, the classification process is carried out. Two distinct feature extraction methods are applied to microarray gene expression data analysis: ranking-based and set-based feature extractions. To enhance extraction performance, a Multi Algorithm Fusion (MAF) method is introduced. Subsequently, the Polynomial Support Vector Machine (PSVM) integrated with MAF is employed for feature extraction. The absolute weights of SVM, fisher ratio, and PSVM are determined, and an MAF-based PSVM algorithm is utilized to achieve efficient results. The proposed MAF-PSVM method demonstrates superior feature robustness compared to traditional approaches, resulting in a significantly lower classification error. Finally, a Random Forest (RF) classifier is utilized for efficient classification, validated by metrics such as error rate, accuracy, specificity, precision, recall, F1 score, and false positive rate.