Table 4 MCC and accuracy values obtained from RF and SVM for testing data for different RFE feature subsets
From: BaPreS: a software tool for predicting bacteriocins using an optimal set of features
Feature set after RFE | Machine learning models | \(Test_{MCC}\) | \(Test_{Acc}\) | Confidence interval |
|---|---|---|---|---|
RFE-MDG-RF | RF | 0.8763 | 0.9464 | (0.887, 0.9801) |
RFE-MDG-SVM | RF | 0.8934 | 0.9464 | (0.887, 0.9801) |
RFE-MDG-RF | SVM | 0.8219 | 0.9107 | (0.8419, 0.9564) |
RFE-MDG-SVM | SVM | 0.8219 | 0.9107 | (0.8419, 0.9564) |
RFE-t-test-RF | RF | 0.8763 | 0.9375 | (0.8755, 0.9745) |
RFE-t-test-SVM | RF | 0.8593 | 0.9286 | (0.8641, 0.9687) |
RFE-t-test-RF | SVM | 0.7862 | 0.8929 | (0.8203, 0.9434) |
RFE-t-test-SVM | SVM | 0.9109 | 0.9554 | (0.8989, 0.9853) |