Improved nu-support vector regression algorithm based on principal component analysis
Principal component analysis (PCA) is the most commonly used approach for analysing high-dimensional data in order to achieve dimension reduction. However, outliers have an adverse effect on the PCA, and hence reduce the accuracy of the prediction model. To date, no research has been done to incorpo...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Published: |
Editura Academia de Studii Economice
2023
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/108931/ |
| Summary: | Principal component analysis (PCA) is the most commonly used approach for analysing high-dimensional data in order to achieve dimension reduction. However, outliers have an adverse effect on the PCA, and hence reduce the accuracy of the prediction model. To date, no research has been done to incorporate the PCA into the algorithm of support vector regression (SVR) technique in order to obtain an accurate prediction model with high accuracy. This paper focuses on improving the nu-SVR algorithm to handle the problem of outliers. A new hybrid PCA with the nu-SVR technique (PCA-SVR) has been established. The performance of the proposed PCA-SVR algorithm is extensively assessed by two real data sets and simulation studies. The outcomes indicate that the PCA-SVR algorithm is more efficient and reliable than the nu-SVR. |
|---|