Help our model "Get Good" (or: "Get Good (enough)")
Closed this issue · 1 comments
ericpan64 commented
With the initial infrastructure set-up for all parts of the project, we've been able get the features generated (average counts of concept_ids), perform some initial filtering using Pearson Correlation/PCA, run the data through the model selection framework (LR, SVM, Random Forest), and use NLP to identify concept_ids that are promising based on the separate CORD dataset. Good stuff!
However, our model still needs to "get good". Let's evaluate using framework below, feel free to add/expand/modify as you see fit. Add new posts with major updates (I can organize this during meetings)
Current best:
| Local Test Results | DREAM Test Results | |
|---|---|---|
| AUC | 0.66 | ... |
| AUPR | ... | ... |
| Balanced Accuracy | ... | ... |
| Features Used | 464 clinical-only features | ... |
| Best Model | LR | ... |
| Other Notes | Pulled info from paper submission | ... |
ericpan64 commented
Model is "good-enough" for the report. Good work everyone!