本文旨在探索支持向量机(SVM)在棋盘数据分类中的应用效果及其性能,特别是在国际象棋和围棋等棋类游戏的局面分类问题上。通过对不同参数设置下的SVM模型进行实验,本文分析了线性核、多项式核及径向基函数(RBF)核SVM在处理高维、复杂棋...本文旨在探索支持向量机(SVM)在棋盘数据分类中的应用效果及其性能,特别是在国际象棋和围棋等棋类游戏的局面分类问题上。通过对不同参数设置下的SVM模型进行实验,本文分析了线性核、多项式核及径向基函数(RBF)核SVM在处理高维、复杂棋局数据时的准确率和泛化能力。本文对比了多种SVM模型在棋盘数据上的分类性能,通过交叉验证和细致的参数调优过程,选出了最优模型。实验结果表明,SVM模型尤其是采用RBF核的模型,在棋盘数据分类任务中展示出了显著的性能优势,包括高准确率和良好的泛化能力。此外,实验也揭示了特征选择和模型参数调优在提高分类性能中的重要性。This paper aims to explore the application effect and performance of support vector machine (SVM) in chessboard data classification, especially in the situation classification of chess and go. Through experiments on SVM models with different parameter settings, the accuracy and generalization ability of linear kernel, polynomial kernel and radial basis function (RBF) kernel SVM in processing high-dimensional and complex chess data are analyzed in this study. In this paper, the classification performance of multiple SVM models on chessboard data is compared, and the optimal model is selected through cross-validation and meticulous parameter tuning process. The experimental results show that the SVM model, especially the model with RBF kernel, shows significant performance advantages in chessboard data classification tasks, including high accuracy and good generalization ability. In addition, the experiment also reveals the importance of feature selection and model parameter tuning in improving classification performance.展开更多
滑坡灾害易发性分析评价对地质灾害的防治与管理具有重要意义。针对滑坡灾害样本选择策略,单核支持向量机多特征映射不合理的问题,本文提出顾及样本优化选择的多核支持向量机(multiple kernel support vector machine,MKSVM)滑坡灾害易...滑坡灾害易发性分析评价对地质灾害的防治与管理具有重要意义。针对滑坡灾害样本选择策略,单核支持向量机多特征映射不合理的问题,本文提出顾及样本优化选择的多核支持向量机(multiple kernel support vector machine,MKSVM)滑坡灾害易发性分析评价方法。为了保证样本平衡性并提高负样本的合理性,采用相对频率比(relative frequency,RF)综合评价各状态对于滑坡灾害易发性影响的重要程度,实现各评价因子状态的合理划分;利用确定性系数法(certainty factor,CF)计算各评价因子各状态分级影响滑坡灾害的敏感性,并在此基础上进行加权求和得到各栅格单元的滑坡灾害易发性指数,在滑坡灾害易发性指数极低和低易发区内随机选择与滑坡灾害点数目一致的非滑坡灾害点作为负样本数据。利用MKSVM对各特征空间最优核函数进行线性组合,解决了单一核函数映射不合理的问题,提高了模型的分类准确率和预测精度。以湖南省湘西土家族苗族自治州为研究区,从滑坡灾害易发性分区图、分区统计及评价模型精度3个方面对CF样本策略的MKSVM模型、CF样本策略的单核SVM模型、随机样本策略的MKSVM模型、随机样本策略的单核SVM模型进行了对比分析。结果表明,4种模型的受试者工作特征曲线(receiver operating characteristic,ROC)下的面积(area under curve,AUC)分别为0.859、0.809、0.798、0.766,验证了CF样本策略的合理性、有效性及MKSVM模型的可靠性。展开更多
文摘本文旨在探索支持向量机(SVM)在棋盘数据分类中的应用效果及其性能,特别是在国际象棋和围棋等棋类游戏的局面分类问题上。通过对不同参数设置下的SVM模型进行实验,本文分析了线性核、多项式核及径向基函数(RBF)核SVM在处理高维、复杂棋局数据时的准确率和泛化能力。本文对比了多种SVM模型在棋盘数据上的分类性能,通过交叉验证和细致的参数调优过程,选出了最优模型。实验结果表明,SVM模型尤其是采用RBF核的模型,在棋盘数据分类任务中展示出了显著的性能优势,包括高准确率和良好的泛化能力。此外,实验也揭示了特征选择和模型参数调优在提高分类性能中的重要性。This paper aims to explore the application effect and performance of support vector machine (SVM) in chessboard data classification, especially in the situation classification of chess and go. Through experiments on SVM models with different parameter settings, the accuracy and generalization ability of linear kernel, polynomial kernel and radial basis function (RBF) kernel SVM in processing high-dimensional and complex chess data are analyzed in this study. In this paper, the classification performance of multiple SVM models on chessboard data is compared, and the optimal model is selected through cross-validation and meticulous parameter tuning process. The experimental results show that the SVM model, especially the model with RBF kernel, shows significant performance advantages in chessboard data classification tasks, including high accuracy and good generalization ability. In addition, the experiment also reveals the importance of feature selection and model parameter tuning in improving classification performance.
