CASE STUDY

Reading the Molecular Labels in Cancer

Reading the Molecular Labels in Cancer

We uploaded expression data from 109 cancer samples with representation from 14 different tumor types onto the JADBio platform to illustrate how automated machine learning can lead to successful predictions but also enhance knowledge discovery.
Read more: Multiclass Cancer Diagnosis

JADBio produced two models, a Best Performing Model, with an AUC of .896 and a Best Interpretable model with an AUC of .679 in two minutes hands-on time and nine hours of analysis time. In addition to the models, the resulting feature selection and visualizations reveal key information about the individual cancer types and the expression data that differentiates their classification.

OTHER