Module 2

Symptom-based herbal recommendation

Type how you feel in everyday words. The analysis service transforms your text with a TF-IDF vectorizer fitted on the curated symptom, disease and remedy datasets, ranks records by cosine similarity, and explains every match.

Live TF-IDF matching

Describe your symptoms

For example: “dry cough with a sore throat for two days”.

Example inputs

How matching works

  1. Your text is analysed with the same tokenizer used to fit the vectorizer
  2. It becomes a TF-IDF vector over the fitted vocabulary
  3. Cosine similarity is computed against every stored record
  4. The highest scoring records above the similarity threshold are returned

API call: POST /api/recommend

Recommendations

Nothing matched yet

Describe your symptoms or pick an example to see up to three herbal records with similarity scores and the reason for each match.