Fig. 1

Pre-screening a variety of ML models with BenchmarkDR. (a) Classification performance comparison of the cross-validated balanced mean accuracy results from the pre-screening. The colour scale describes the model’s performance from red (good) to blue (poor). (b) Regression performance comparison of the cross-validated average mean squared error results from the pre-screening. The colour scale describes the model’s performance from red (good) to blue (poor). The darker shades (right columns) correspond to the regression of non-transformed MIC values, and the lighter shades (left columns) correspond to the regression of log₂ transformed MIC values.