基于网格搜索超参数优化的支持向量回归

Support Vector Regression Based on Grid Search Hyperparameter Optimization

支持向量机在数据挖掘中有良好表现并成功应用于诸多领域,但是其超参数对模型精度影响显著。为了获得更好的回归预测结果,本文使用网格搜索算法进行超参数优化,通过比较三种评价指标和可视化对比实验,最后获得具有较小误差的支持向量回归机,同时为支持向量机的超参数优化问题提供参考。

Support vector machine has a good performance in data mining and has been successfully applied in many fields,but its hyperparameters have a significant impact on model accuracy.In order to obtain better regression prediction results,this paper uses grid search algorithm to optimize hyperparameters.By comparing three evaluation indicators and visual comparison experiments,a support vector regression machine with smaller error is finally obtained.The parameter optimization problem provides a reference.