Article

Evaluating Supervised Predictive Models for Automated Diagnosis of Agrochemical Poisoning

Author : M VINOD KUMAR, K PRAVEEN KUMAR, K SOWMYA, P V KOMALI

In a Data Science project, it is essential to assess data relevance and identify patterns that support decisionmaking through domain-specific knowledge. Clearly defining methodologies and creating documentation to guide a project from inception to completion are also vital. This study introduces a Data Science model that guides the process from data collection through model training to facilitate knowledge discovery. The model was motivated by shortcomings in existing Data Science methodologies, especially the lack of practical, stepby-step guidance for preparing data for production. Called the “Data Refinement Cycle with Supervised Machine Learning (DRC–SML),” the proposed model was developed to meet the needs of a Data Science project designed to help healthcare professionals diagnose pesticide poisoning in rural workers. The project dataset came from scientific research involving 1,027 samples with data on toxicity biomarkers and clinical analyses. The model achieved 99.61% accuracy using just 27 rules to determine diagnoses. These results optimized healthcare practices and improved quality of life in rural areas, demonstrating the success of the proposed model.


Full Text Attachment
//