Microarray Data Classification Accuracy Enhancement Using Support Vector Macine in Comparison with K-Nearest Neighbour

The goal of this study was to evaluate the effectiveness of machine learning approaches for classifying microarray data by comparing the SVM and KNN algorithms. This work involved using genetic microarray data and analyzing the classification performance of SVM and KNN classifiers. For the analysis, G-power calculations were set with an 80 % power level, a 95 % confidence interval, and alpha and beta values of 0.05 and 0.2, respectively, with a total sample size of 20. The results showed that the KNN algorithm achieved an accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 6. 6 6 \%}$</tex> and a loss rate of 3.33%, outperforming the Support Vector Machine, which had an accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 1. 3 1 \%}$</tex> and a loss rate of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8. 6 8 \%}$</tex>. Despite these differences, the statistical analysis using an independent sample <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$t$</tex>-test indicated that the performance differences between SVM and KNN were not statistically significant, with a p-value of 0.10 at a 95% confidence level. Overall, the KNN algorithm demonstrated a notably higher classification accuracy compared to the Support Vector Machine, making it a more effective choice for microarray data classification.