基于网格搜索优化支持向量机多分类参数识别不同工艺酱酒的应用研究

Application of grid search-optimized support vector machine multi-classification parameters inidentifying sauce-flavor Baijiu with different processes

为提升支持向量机(SVM)在不同工艺酱酒分类预测中的准确度,该实验利用网格搜索优化支持向量机参数,建立最优参数的支持向量机分类预测模型。通过对不同工艺酱香型白酒客观结构特征定量分析,将提取的特征信息数据经过预处理(异常值处理、归一化操作等)后存储为样本数据集。其中样本数据分为训练样本与测试样本,通过训练样本对最优参数的SVM白酒品牌分类预测模型进行训练,测试样本对模型进行预测分类。经过试验验证,该模型的不同工艺分类识别率达到94.44%,较传统的SVM等分类算法能够快速、有效地对不同工艺的酱酒进行分类识别,显著改善分类的精度,改进后的方法实现过程也比较简单。

To enhance the accuracy of support vector machine(SVM)in the classification and prediction of sauce-flavor(Jiangxiangxing)Baijiu with different processes,a grid search was employed to optimize the parameters of the SVM,and SVM classification prediction model with the optimal parameters was established.Through quantitative analysis of the objective structural characteristics of sauce-flavor Baijiu with different processes,the extracted feature information data was preprocessed(including outlier handling and normalization)and stored as a sample dataset.The sample data were divided into training samples and testing samples.The training samples were used to train the SVM Baijiu brand classification prediction model with optimal parameters,and the test samples were used to predict the classification of the testing samples.The experimental verification demonstrated that the classification recognition rate of the model for different processing technologies reached 94.44%.Compared with the traditional classification algorithms such as SVM,the model could quickly and effectively classify and recognize sauce-flavor Baijiu with different processes,and significantly improving classification accuracy.The improved method was also relatively simple in implementation.