Answers are grounded in this paper and cite their sources.
The recent revival in the popularity of Ayurvedic medicine demands the smart digital systems which will be able to prescribe medicinal herbs according to the individualized symptom picture. In this paper, the proposed method provides a lightweight and explainable hybrid model, termed as MedLeafRec, which can suggest the Ayurvedic medicinal leaves and their suitable dosage, given the input features, i.e., age, gender, type of symptom, temperature, and severity. MedLeafRec incorporates a two-level decision-making method: rule-based inference engine, which relies on Ayurvedic expertise, and a fallback decision tree classifier, which deals with the situations that were not covered by predefined mappings. Prediction of dosage is achieved by using a linear regression model that incorporates the use of normalized physiological parameters to predict quantity in either grams or milliliters. Comprehensive testing on a selected dataset proves that MedLeafRec has a dosage prediction Mean Absolute Error (MAE) of 0.62 g/ml and a classification accuracy of 95.34%. Such performances are substantially higher than those of baseline models, such as Random Forest (89.45%), SVM (87.50%), and Rule-Only Systems (82.35%). In addition, the model has a small footprint (2.1 MB) and low inference latency (3.4 ms/sample), which makes it very applicable in mobile and constrained settings. The modular and transparent design of MedLeafRec allows it to integrate with healthcare platforms that can be deployed in the field without disturbing the clinical reasoning of the conventional practice.
No takes yet. Share an insight, caveat, or question.
Madula et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: