New approach slashes errors in regression models, boosts anti-noise capability.
The article presents a new method to improve support vector regression by confirming weight values based on linear programming. This approach helps reduce errors caused by noise and outliers in the data. By assigning weight values to input samples based on their distance to the database, the method enhances the accuracy of regression and boosts the ability to handle noisy data. Experimental results demonstrate that this approach effectively reduces regression errors and enhances the anti-noise capability of support vector machines.
