Predictive Study on the Occurrence of Wheat Blossom Midges Based on Gene Expression Programming with Support Vector...

Area of Science:

• Agricultural Science
• Computational Biology
• Machine Learning

Background:

• Smart agriculture faces challenges in efficient plant pest and disease prediction.
• Existing models often suffer from slow training speeds and low prediction accuracy.
• There is a need for advanced computational methods to improve predictive capabilities.

Purpose of the Study:

• To introduce an innovative prediction method integrating gene expression programming (GEP) with support vector machines (SVM) for enhanced pest prediction.
• To address the limitations of existing models in terms of speed and accuracy.
• To optimize agricultural production strategies through improved predictive modeling.

Main Methods:

• Developed the gene expression programming-support vector machine (GEP-SVM) model.
• Utilized individual gene values as parameters for SVM, optimized via grid search for genetic parameters.
• Tested the model on historical wheat blossom midge data (1933-2010) and compared it with traditional methods.

Main Results:

• The GEP-SVM model achieved a back-generation accuracy rate of 90.83%.
• Demonstrated superior generalization and fitting capabilities compared to GEP, SVM, naive Bayes, K-nearest neighbor, and BP neural networks.
• Significantly enhanced computational efficiency in pest and disease prediction.

Conclusions:

• The GEP-SVM model offers a significant advancement in smart agriculture for pest and disease prediction.
• The model provides a robust scientific foundation for future predictive endeavors in agriculture.
• This approach contributes to the optimization of agricultural production strategies through enhanced predictive accuracy and efficiency.