The Additive Input-Doubling Method Based on the SVR with Nonlinear Kernels: Small Data Approach

The study introduces an additive input-doubling method designed for processing short and very short datasets, addressing challenges in fields such as medicine, economics, and materials science. Traditional machine learning models struggle with limited data, often leading to unreliable predictions. To overcome this issue, the authors propose a novel data augmentation technique that expands datasets both in rows and columns without requiring additional training. The method leverages Support Vector Regression (SVR) with nonlinear kernels to enhance prediction accuracy. The approach follows principles of axial symmetry and has been experimentally validated, outperforming existing methods in handling small datasets. Two algorithmic implementations of the method are presented, with optimized operational parameters. The research also discusses potential applications, limitations, and future improvements of the method.

Takeaways:

1. Small data processing: The study introduces an additive input-doubling method for enhancing the analysis of short and very short datasets.
2. Data augmentation technique: The proposed method expands datasets in both rows and columns without requiring additional training.
3. SVR with nonlinear kernels: The method utilizes Support Vector Regression (SVR) with nonlinear kernels to improve prediction accuracy.
4. Axial symmetry principle: The data augmentation approach aligns with axial symmetry principles, ensuring consistent dataset expansion.
5. Algorithmic implementations: Two implementations of the method are presented, with optimal parameters selected through experimentation.
6. Performance comparison: The method was experimentally validated and demonstrated superior prediction accuracy compared to existing approaches.
7. Application areas: The approach is applicable in medicine, economics, materials science, and other domains where small datasets pose a challenge.
8. Future research: The study highlights limitations and outlines prospects for improving the method’s efficiency and expanding its use cases.