模型

测试集分类正确率

测试集召回率

RF-Adaboost

84.62%

80.40%

LSTM神经网络

81.19%

78.00%

RBF-SVM

74.36%

66.68%

Linear-SVM

75.21%

71.60%

基于核密度估计的朴素贝叶斯

78.64%

37.55%