Missing value handling

---

Preface

In daily statistical analysis , Missing data is a common phenomenon , The data may be missing in a data entry , Or you don't get the result you want in the Q & A , Will cause partial loss of data . There are many ways to deal with missing data , We need to solve it according to the actual situation . When the proportion of missing data in the whole is relatively small , We can delete , But if the data has a significant impact on the results , Then we cannot delete , The data needs to be filled .

---

1、 Detection of missing values ？

stay R In language , The sign of our missing value is NA To represent the , And in R in , Use functions is.na() It can be used to detect whether there are missing values in the data .

na <- c(2,2,NA,6,8)
is.na(na)

result ：

Use is.na() Function to determine if there is a missing value , The returned result is a logical value , That is, with or without missing values . When the returned logical value is FALSE There is no missing value , When the return value is TURE Is the missing value .

---

2、NaN Detection of

NaN Represents a value that cannot be represented , The value of uncertainty , It can be used is.nan() To detect

nan <- c(5,Inf,5,NaN,6)
is.nan(nan)

result

---

3、 Processing of missing values

Borrow and reissue ：

new_na <- c(1,1,2,3,NA,7,8)
new_na[is.na(new_na)] <- mean(new_na,na.rm=T)

result

---

4、 Select a number that is not a missing value

Add a missing value in the data , Then we cannot calculate , So we need to filter the missing values to calculate . Then we just need to select a number that is not a missing value .

new_nas <- c(1,NA,1,2,3,NA,7,8)
new_nas <- new_nas[!is.na(new_nas)]

result ：

---

5、 Matrix with missing values

The matrix can always have missing values , But the same can not be calculated , We need to deal with the missing values before we can calculate .

mat <- matrix(c(1:11,NA),nr=3,nc=4,byrow = T)

result

---

6、 Not a missing value

mean(mat,na.rm=T)

result

---

7、 Delete the line

When we delete NA The row where the value is located , There are no missing values in the matrix , So this matrix can be calculated .

newmat <- na.omit(mat)

result

---

8、 Calculating mean

We find a mean value of the number in the matrix , You can get the following results .

mean(newmat)

result

---

summary

1：is.na() Missing values can be detected , And return the logical vector ;
2：is.nan() It can detect NaN value , And return the logical vector ;
3： Those with missing values cannot be calculated .