R：机器学习实战
参考：http://blog.csdn.net/liuxincumt/article/details/7527917
1 神经网络
nnet包
# # 示例
1
ir <- rbind(iris3[,,
1
],iris3[,,
2
],iris3[,,
3
])
targets <-
class
.ind( c(rep(
"s"
,
50
), rep(
"c"
,
50
), rep(
"v"
,
50
)) )
samp <- c(sample(
1
:
50
,
25
), sample(
51
:
100
,
25
), sample(
101
:
150
,
25
))
ir1 <- nnet(ir[samp,], targets[samp,], size =
2
, rang =
0.1
,
decay = 5e-
4
, maxit =
200
)
test.cl <- function(
true
, pred) {
// 测试函数
true
<- max.col(
true
)
cres <- max.col(pred)
table(
true
, cres)
}
test.cl(targets[-samp,], predict(ir1, ir[-samp,]))
# 示例
2
ird <- data.frame(rbind(iris3[,,
1
], iris3[,,
2
], iris3[,,
3
]),
species = factor(c(rep(
"s"
,
50
), rep(
"c"
,
50
), rep(
"v"
,
50
))))
ir.nn2 <- nnet(species ~ ., data = ird, subset = samp, size =
2
, rang =
0.1
,
decay = 5e-
4
, maxit =
200
)
table(ird$species[-samp], predict(ir.nn2, ird[-samp,], type =
"class"
))
|
分类器：
# weights: 19 initial value 82.459767 iter 10 value 27.862946 iter 20 value 7.244425 iter 30 value 3.255549 iter 40 value 3.169972 iter 50 value 2.792178 iter 60 value 1.766152 iter 70 value 1.272564 iter 80 value 0.852201 iter 90 value 0.681699 iter 100 value 0.531267 iter 110 value 0.489096 iter 120 value 0.470457 iter 130 value 0.464765 iter 140 value 0.461259 iter 150 value 0.457745 iter 160 value 0.457484 iter 170 value 0.457441 iter 180 value 0.457428 iter 190 value 0.457425 final value 0.457424 converged
结果：
c s v c 24 0 1 s 0 25 0 v 3 0 22
2 回归
2.1 线性回归树（决策树）：
CART：Classification and Regression Trees
数据集：stagec
- 数目：146
- 属性：pgtime pgstat age eet g2 grade gleason ploidy
R code：
library(rpart)
progstat = factor(stagec$pgstat, levels=
0
:
1
, labels=c(
"No"
,
"Prog"
))
cfit = rpart(progstat~age+eet+g2+grade+gleason+ploidy, data=stagec,method=
"class"
)
print(cfit)
par(mar=rep(
0.1
,
4
))
plot(cfit)
text(cfit)
|
结果：
1) root 146 54 No (0.6301370 0.3698630)
2) grade< 2.5 61 9 No (0.8524590 0.1475410) *
3) grade>=2.5 85 40 Prog (0.4705882 0.5294118)
6) g2< 13.2 40 17 No (0.5750000 0.4250000)
12) ploidy=diploid,tetraploid 31 11 No (0.6451613 0.3548387)
24) g2>=11.845 7 1 No (0.8571429 0.1428571) *
25) g2< 11.845 24 10 No (0.5833333 0.4166667)
50) g2< 11.005 17 5 No (0.7058824 0.2941176) *
51) g2>=11.005 7 2 Prog (0.2857143 0.7142857) *
13) ploidy=aneuploid 9 3 Prog (0.3333333 0.6666667) *
7) g2>=13.2 45 17 Prog (0.3777778 0.6222222)
14) g2>=17.91 22 8 No (0.6363636 0.3636364)
28) age>=62.5 15 4 No (0.7333333 0.2666667) *
29) age< 62.5 7 3 Prog (0.4285714 0.5714286) *
15) g2< 17.91 23 3 Prog (0.1304348 0.8695652) *
2.2 随机森林
随机森林是一个包含多个决策树的分类器， 并且其输出的类别是由个别树输出的类别的众数而定。（http://baike.baidu.com/view/5021113.htm）
randomForest包官网：http://www.stat.berkeley.edu/~breiman/RandomForests/cc_home.htm
2 SVM
e1071包
library(
"e1071"
)
model <- svm(Species ~ ., data = iris,
method =
"C-classification"
,
kernel =
"radial"
,
cost =
10
, gamma =
0.1
)
summary(model)
par(mar=rep(
0
,
4
))
plot(model, iris,
Petal.Width ~ Petal.Length,
slice = list(Sepal.Width =
3
,Sepal.Length =
4
))
|