What it involved
- Classifier trained on 147,000 UCI drug reviews, scoring 90 percent against 66 percent for an NLTK baseline.
- A Flask service scores reviews and writes them to MongoDB, backfilled in stages rather than one pass.
- Validates and normalizes inconsistent drug, dosage, and interaction data before the model sees it.
- Validation exercised against malformed and out-of-range inputs.
This one does not have a full write-up yet. If you want the detail on it, ask me directly and I will walk you through the decisions.