机器学习基于R(2)--分类--逻辑回归

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19

data(titanic_train, package = "titanic")
titanicSub = titanic_train[,c("Survived","Sex","Pclass",
                                "Age","Fare","SibSp","Parch")]
summary(titanicSub)
# Survived          Sex                Pclass           Age             Fare            SibSp           Parch       
# Min.   :0.0000   Length:891         Min.   :1.000   Min.   : 0.42   Min.   :  0.00   Min.   :0.000   Min.   :0.0000  
# 1st Qu.:0.0000   Class :character   1st Qu.:2.000   1st Qu.:20.12   1st Qu.:  7.91   1st Qu.:0.000   1st Qu.:0.0000  
# Median :0.0000   Mode  :character   Median :3.000   Median :28.00   Median : 14.45   Median :0.000   Median :0.0000  
# Mean   :0.3838                      Mean   :2.309   Mean   :29.70   Mean   : 32.20   Mean   :0.523   Mean   :0.3816  
# 3rd Qu.:1.0000                      3rd Qu.:3.000   3rd Qu.:38.00   3rd Qu.: 31.00   3rd Qu.:1.000   3rd Qu.:0.0000  
# Max.   :1.0000                      Max.   :3.000   Max.   :80.00   Max.   :512.33   Max.   :8.000   Max.   :6.0000

# 第一列：生存与否0/1
# 第二列：性别
# 第三列：头等舱、二等舱、三等舱 1/2/3
# 第四列：年龄
# 第五列：票价
# 第六列：兄弟姐妹+配偶人数
# 第七列：父母和孩子总人数

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20

#（1）分类变量因子化
#（2）挑选合适特征（第六列+第七列，表示家庭成员数）
fctrs <- c("Survived", "Sex", "Pclass")
titanicClean <- titanicSub %>%
  mutate_at(.vars = fctrs, .funs = factor) %>%
  mutate(FamSize = SibSp + Parch) %>%
  select(Survived, Pclass, Sex, Age, Fare, FamSize)
head(titanicClean)
#   Survived Pclass    Sex Age    Fare FamSize
# 1        0      3   male  22  7.2500       1
# 2        1      1 female  38 71.2833       1
# 3        1      3 female  26  7.9250       0
# 4        1      1 female  35 53.1000       1
# 5        0      3   male  35  8.0500       0
# 6        0      3   male  NA  8.4583       0

# （3） 在年龄列有缺失值（NA），需要处理
##如下均值用缺失值代替
imp <- impute(titanicClean, cols = list(Age = imputeMean()))
sum(is.na(imp$data$Age)) #0

1
2

titanicTask <- makeClassifTask(data = imp$data, target = "Survived")
titanicTask

1
2

#设置predict.type参数为"prob"，则预测输出不仅仅是分类变量，还有概率值
logReg <- makeLearner("classif.logreg", predict.type = "prob")

1
2

#训练模型
logRegModel <- train(logReg, titanicTask)

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18

logRegModelData <- getLearnerModel(logRegModel)
coef(logRegModelData)
# (Intercept)      Pclass2      Pclass3      Sexmale          Age         Fare      FamSize 
# 3.809661697 -1.000344806 -2.132428850 -2.775928255 -0.038822458  0.003218432 -0.243029114

#指数转换
exp(cbind(Odds_Ratio = coef(logRegModelData), confint(logRegModelData)))
#              Odds_Ratio       2.5 %       97.5 %
# (Intercept) 45.13516691 19.14718874 109.72483921
# Pclass2      0.36775262  0.20650392   0.65220841
# Pclass3      0.11854901  0.06700311   0.20885220
# Sexmale      0.06229163  0.04182164   0.09116657
# Age          0.96192148  0.94700049   0.97652950
# Fare         1.00322362  0.99872001   1.00863263
# FamSize      0.78424868  0.68315465   0.89110044

#对于连续变量的解释：例如Age：表示其它变量不变，每Age增长1岁，生存率降低4%
#对于分类变量的解释，需要有参照：例如Sexmale：表示男性的生存率仅为女性的6%

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

data(titanic_test, package = "titanic")
titanicNewClean <- titanic_test %>%
  mutate_at(.vars = c("Sex", "Pclass"), .funs = factor) %>%
  mutate(FamSize = SibSp + Parch) %>%
  select(Pclass, Sex, Age, Fare, FamSize)

LogiPred = predict(logRegModel, newdata = titanicNewClean)
head(LogiPred$data)
#      prob.0     prob.1 response
# 1 0.9178036 0.08219636        0
# 2 0.5909570 0.40904305        0
# 3 0.9123303 0.08766974        0
# 4 0.8927383 0.10726167        0
# 5 0.4069407 0.59305933        1
# 6 0.8337609 0.16623907        0

#由于在这个test测试集里没有Survived列，所以返回的对象列没有表示真实生存状态的truth列

1
2

kFold <- makeResampleDesc(method = "RepCV", folds = 10, reps = 50, 
                          stratify = TRUE)

1
2
3

#由于在数据预处理步骤时，使用均值模拟了缺失值。在交叉验证时也需要将这一步纳入进去
logRegWrapper <- makeImputeWrapper("classif.logreg",
                                   cols = list(Age = imputeMean()))

1
2
3
4
5
6
7
8

logRegwithImpute <- resample(logRegWrapper, titanicTask, resampling = kFold, 
                             measures = list(acc, fpr, fnr))
logRegwithImpute
# Resample Result
# Task: imp$data
# Learner: classif.logreg.imputed
# Aggr perf: acc.test.mean=0.7965652,fpr.test.mean=0.2987697,fnr.test.mean=0.1440801
# Runtime: 8.90803