This project explores the fusion of Ayurveda and machine learning to classify human Prakriti (body constitution) types into Vata, Pitta, and Kapha. The project performed excellent to Random Forest-based ensemble model to achieve high accuracy, providing a scientific method for personalized health insights based on Ayurveda. A detailed project PDF has been created and uploaded to the repository, covering the project’s entire flow, including methodology, feature extraction, and results analysis.
Explanation of the Random Forest Prakriti Classification (RFPC) method, along with hyperparameter tuning and ensemble learning strategies. A methodology diagram is provided below to illustrate this process.
Achieved 98% accuracy using Random Forest. The results analysis table below presents performance metrics (precision, recall, F1-score) for each Prakriti class.
Our research paper based on this project has been successfully published in the IEEE Xplore proceedings of the IACIS 2024 conference. This publication highlights the use of ensemble learning models for predicting human body Prakriti using Ayurvedic Tridosha features, providing a scientific basis for integrating machine learning with traditional Ayurvedic principles.
- Title: Predicting Human Body Prakriti with Ayurvedic Tridosha Features Using Ensemble Learning Models
- Conference: International Association for Computer Information Systems (IACIS) 2024
- Date Added to IEEE Xplore: 24 October 2024
- DOI: 10.1109/IACIS61494.2024.10721947
- Link: https://ieeexplore.ieee.org/document/10721947
A. Bhosale, P. Girase, A. Waghmare, S. Barve and D. Chikmurge, "Predicting Human Body Prakriti with Ayurvedic Tridosha Features using Ensemble Learning Models," 2024 International Conference on Intelligent Algorithms for Computational Intelligence Systems (IACIS), Hassan, India, 2024, pp. 1-7,
Ayurveda, an ancient medical system, provides insight into personal health through the concept of Prakriti, which determines a person's constitution. Integrating these traditional concepts with modern medical practice is a huge challenge. The research has addressed this problem by developing a machine learning-based Prakriti estimation method. Advanced algorithms, including random forests and ensemble models, were used to analyze a rich dataset of biological traits not easily altered by current cosmetic or technological interventions. This approach achieved a remarkable accuracy of 98%, validating the effectiveness of integrating machine learning and Ayurvedic principles. Future research will focus on expanding the dataset to include a wider range of biological features and improving feature selection methods to increase the generalizability and applicability of the model in various clinical settings, thereby improving the relationship between traditional Ayurvedic knowledge and modern medical practice. The gap must be narrowed further. keywords: {Accuracy;Machine learning algorithms;Biological system modeling;Estimation;Medical services;Predictive models;Feature extraction;Random forests;IEEE Constitution;Faces;machine learning;health domain;ayurveda;prakriti;prakriti identification;prashna pariksha},URL: https://ieeexplore.ieee.org/document/10721947