Analysis and Identification of Aptamer-Compound Interactions with a Maximum Relevance Minimum Redundancy and Nearest Neighbor Algorithm

Figure 2

Four IFS curves plotted by taking MCC as the -axis and the number of considered features as the -axis for four basic prediction engines. The MCC values indicate the performance of various prediction models using different classifiers and different combination of features to represent interactions. It can be observed that using NNA as the classifier and the first 80 features in the mRMR features list to represent interactions can yield the best performance with the highest MCC value of 0.670.