Skip to main content
Fig. 2 | Biomarker Research

Fig. 2

From: DeepMPTB: a vaginal microbiome-based deep neural network as artificial intelligence strategy for efficient preterm birth prediction

Fig. 2

Performance of DeepMPTB based on the 20% test set (239 samples). A During model evaluation, the three metadata features (age, ethnicity and trimester of sample collection) were associated with each sample. For each sample, the true and predictive phenotypes were compared to evaluate the performance of DeepMPTB. B A summary plot for the SHAP values was generated to understand the contributions of the first 20 features in this performance analysis. Features related to clinical/demographic metadata are indicated by an asterisk. Each dot represents a sample. Negative and positive SHAP values are associated to TB and PTB prediction, respectively. Low and high SHAP values are shown in blue and red, respectively. TB: Term birth; PTB: Preterm birth

Back to article page

Biomarker Research

ISSN: 2050-7771

Contact us