DAPLSR: Data Augmentation Partial Least Squares Regression Model via Manifold Optimization
This work addresses data imbalance issues in PLSR for domain-specific applications, representing an incremental improvement.
The paper tackled the problem of Partial Least Squares Regression (PLSR) underperforming with uneven category data by proposing a DAPLSR model using SMOTE and VDM for data augmentation and manifold optimization for improved numerical solutions, achieving superior classification performance and evaluation metrics on various datasets.
Traditional Partial Least Squares Regression (PLSR) models frequently underperform when handling data characterized by uneven categories. To address the issue, this paper proposes a Data Augmentation Partial Least Squares Regression (DAPLSR) model via manifold optimization. The DAPLSR model introduces the Synthetic Minority Over-sampling Technique (SMOTE) to increase the number of samples and utilizes the Value Difference Metric (VDM) to select the nearest neighbor samples that closely resemble the original samples for generating synthetic samples. In solving the model, in order to obtain a more accurate numerical solution for PLSR, this paper proposes a manifold optimization method that uses the geometric properties of the constraint space to improve model degradation and optimization. Comprehensive experiments show that the proposed DAPLSR model achieves superior classification performance and outstanding evaluation metrics on various datasets, significantly outperforming existing methods.