Some classes of problem are best solved by a specific class of machine learning model, due to the structure of the data (e.g. CNN's for computer vision tasks).

Prediction of bacterial resistance/susceptibility to antimicrobials (from genotypic data) using Machine Learning methods is a problem that has started receiving interest in recent years.

The following paper (from 2017) analysed the then current literature and found that:

To date, there has not been a consensus about the optimal machine learning model to be used for AST genotype–phenotype prediction, as reflected by the diverse algorithms authors have implemented (Table 1).

Has this changed in the past 2 years?

Is there now a consensus about which models are most effective?

References: Table 1 from paper:

enter image description here



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.