给你一个例子：
准备数据
1、data：一个表达谱数据矩阵，行为样本，列为基因
2、risk：一个向量，各样本的分类向量
library('e1071')
library(ROCR)
train_data
train_label
train_label
obj
cost = 2^(0:2)),
tunecontrol = tune.control(sampling = "cross"))##找最优的gamma，和cost
best_gm
best_ct
model
kernel = "radial", gamma = best_gm, cost = best_ct, cross = 0)
pre_label
stat_res
accurary
sensitivity
specificity
pre_prob
pred
perf
plot(perf,colorize=TRUE)
auc.tmp
auc.value
AUC
auc.value