R notes mice

1 MICE Algorithm theory part

        MICE(Multiple Imputation by Chained Equations) It is a robust way to deal with missing data in data sets 、 Informative methods . This process passes through a series of iterative prediction models “ fill ”（ Estimate ） Missing data in dataset .

         In each iteration , Each specified variable in the dataset is estimated using other variables in the dataset . Constantly iterating in , Until it converges .

1.1 MICE give an example

          This process continues until all specified variables are interpolated . If there is no convergence , You can run additional iterations , Although usually no more than 5 This iteration is necessary .

         The accuracy of interpolation depends on the information density in the data set . The data set of completely independent variables without correlation will not produce accurate interpolation .

1.2 PMM,Predictive Mean Matching

        MICE You can use what is called predictive mean matching (PMM) To select the value to be estimated . PMM Select a data point from the original non missing data , The predicted value of this data point is close to the predicted value of the missing sample .

         Choose the closest N Data points as candidate values , Select a value randomly from them to complete .

 2 R Language MICE

2.0 Import package

library(magrittr)
library(dplyr)
library(mice)
library(missForest)

2.1 Import data

data(iris)
summary(iris)


#  Sepal.Length    Sepal.Width     Petal.Length  
# Min.   :4.300   Min.   :2.000   Min.   :1.000  
# 1st Qu.:5.100   1st Qu.:2.800   1st Qu.:1.600  
# Median :5.800   Median :3.000   Median :4.350  
# Mean   :5.843   Mean   :3.057   Mean   :3.758  
# 3rd Qu.:6.400   3rd Qu.:3.300   3rd Qu.:5.100  
# Max.   :7.900   Max.   :4.400   Max.   :6.900  

#  Petal.Width          Species  
# Min.   :0.100   setosa    :50  
# 1st Qu.:0.300   versicolor:50  
# Median :1.300   virginica :50  
# Mean   :1.199                  
# 3rd Qu.:1.800                  
# Max.   :2.500  

2.2 Random loss of a certain amount of data

Randomly generated in the data 10% Of Missing value . At the same time  Species  This categorical variable is also removed .

iris_mis <- missForest::prodNA(iris, noNA = 0.1) %>%    select(-Species)

summary(iris_mis)
#  Sepal.Length    Sepal.Width     Petal.Length  
# Min.   :4.300   Min.   :2.000   Min.   :1.000  
# 1st Qu.:5.100   1st Qu.:2.800   1st Qu.:1.500  
# Median :5.800   Median :3.000   Median :4.300  
# Mean   :5.856   Mean   :3.049   Mean   :3.707  
# 3rd Qu.:6.400   3rd Qu.:3.300   3rd Qu.:5.100  
# Max.   :7.900   Max.   :4.200   Max.   :6.900  
# NA's   :16      NA's   :15      NA's   :7      
#  Petal.Width   
# Min.   :0.100  
# 1st Qu.:0.300  
# Median :1.300  
# Mean   :1.201  
# 3rd Qu.:1.800  
# Max.   :2.500  
# NA's   :13     

2.3 Visualize missing data

md.pattern(iris_mis)

         It means Petal.Length altogether 7 individual missing data , Two of them and the second column Petal.Weight Lose data at the same coordinates ; be left over 5 One only loses data at its own coordinates . 

2.4 To complete

imputed_Data <- mice(iris_mis, m=5, maxit = 50, method = 'pmm', seed = 123)

• m = 5 , To generate 5 Group filled data
• maxit = 50, The number of iterations per generation of fill data , Here take 50 Time
• method = ‘pmm’, Use 1.2 To introduce the Predictive Mean Matching Methods （ Continuous data is used ）

2.5 View the data

Because we generated 5 Group data , So you can check it in groups

completeData <- mice::complete(imputed_Data,2)

This one at the back 2 It means which group of data to view