← Back to Portfolio
🌱 Healthcare AI · Rural Impact

Pesticide Poisoning
Diagnosis System

An AI-powered diagnostic pipeline detecting pesticide poisoning in rural agricultural workers with 98% accuracy — enabling timely, affordable care.

98%
Accuracy
1,027
Clinical Samples
DRC-SML
Algorithm
Social Impact
Why This Project Exists
🌾
Rural Agricultural Workers
Millions of Indian farmers exposed to toxic pesticides daily with no affordable diagnostic access.
⏱️
Faster Diagnosis
Reduces diagnosis time from days to seconds — critical in poisoning emergencies.
💰
Affordable Care
Eliminates need for expensive lab tests — AI inference replaces hours of clinical work.
🏥
Clinic Ready
Designed for low-resource settings with minimal infrastructure requirements.
Core Algorithm
The DRC-SML Approach
Dynamic Rule-based Classification with Supervised ML
Rule-based layer — medical expert rules filter obvious cases and reduce noise first
Supervised ML layer — trained classifier handles complex borderline cases
Dynamic thresholding — adapts classification boundaries based on symptom patterns
Result — 98% accuracy on 1,027 samples, outperforming all standard ML models
Stack
Tech Used
PythonScikit-learnPandas NumPyDRC-SML AlgorithmMatplotlibJupyter Notebook
Developer
Built By
SK
Shiva Keshava
Data Scientist & ML Engineer · B.Tech AI & Data Science (Final Year Project)