Using Machine Learning to Predict Rice Blast Disease
Author Information
Author(s): Kaundal Rakesh, Kapoor Amar S, Raghava Gajendra PS
Primary Institution: Institute of Microbial Technology, Chandigarh, India
Hypothesis
Can support vector machines improve the prediction of rice blast disease severity compared to traditional methods?
Conclusion
The study found that support vector machines significantly outperformed traditional machine learning techniques in predicting rice blast disease severity.
Supporting Evidence
- The support vector machine approach improved the correlation coefficient from 0.50 to 0.77 compared to traditional methods.
- Weather variables such as rainfall were identified as the most influential predictors of rice blast severity.
- The study developed a web-based server for real-time predictions of rice blast severity.
Takeaway
This study shows that a new computer program can help farmers know when to protect their rice plants from a disease called blast, which can hurt their crops.
Methodology
The study used support vector machines to analyze weather data and predict rice blast severity, comparing its performance with multiple regression and neural network approaches.
Limitations
The models may not generalize well across different geographical locations and years due to variability in environmental conditions.
Participant Demographics
Data was collected from five different locations in Himachal Pradesh, India.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website