Gene Expression Analysis Using SVM And KNN Classifiers On Various Datasets

The dimensionality curse occurs due to the high dimensionality of input data, which is crucial in gene expression data analysis. The labeling process is complex with the available dataset and time-consuming. The training process is also complex with machine learning approaches. Various existing approaches intend to enhance interpretability and address the smoothing problem and labeling issues. Here, two diverse classifiers, k-NN (k-Nearest neighbor) and SVM (Support Vector Machine) are used to classify the data and examine the features required for classification. The performance is analyzed using five diverse datasets: BRCA, GSE412, lung, Leukemia and DLBCL to offer practical assistance. Based on the analysis, it is observed that k-NN provides better prediction outcomes compared to SVM and enhances performance. Both these models establish better trade-offs compared to other approaches.