Error while performing linear regression

Question

I am trying to perform Linear Regression on the below data:-

Need to perform Linear Regression on Air_weight and Water_weight.

Kindly let me know how to resolve this error.

This is the code i tried but got an error:-

fit <- lm(Water_Weight~Air_Weight, data=table1)

Error

**Warning messages:
1: In model.response(mf, "numeric") :
  using type = "numeric" with a factor response will be ignored
2: In Ops.factor(y, z$residuals) : ‘-’ not meaningful for factors**



ID  GENDER  Air_Weight    Water_Weight  Body_Fat
01  1       75.60               *        14.17 
02  1       70.70              3.60      13.95 
03  1         *                4.00      8.98 
04  1       95.00              4.30      17.32   
05  1       73.20              3.80      11.50 

Answer 1

You are having some problem with the structure of your data, probably based on the way you read it into R. The most obvious issue is that you will need to use na.strings="*" as an additional argument when reading in your data (with read.csv() or read.table()), to avoid turning the Air_Weight and Water_Weight variables into factors.

There may be other problems, but they are impossible to diagnose remotely. Here's an example that shows this approach can work:

table1 <- read.table(header=TRUE,na.strings="*",text="
ID  GENDER  Air_Weight    Water_Weight  Body_Fat
01  1       75.60               *        14.17 
02  1       70.70              3.60      13.95 
03  1         *                4.00      8.98 
04  1       95.00              4.30      17.32
05  1       73.20              3.80      11.50") 

str(table1)
## 'data.frame':    5 obs. of  5 variables:
##  $ ID          : int  1 2 3 4 5
##  $ GENDER      : int  1 1 1 1 1
##  $ Air_Weight  : num  75.6 70.7 NA 95 73.2
##  $ Water_Weight: num  NA 3.6 4 4.3 3.8
##  $ Body_Fat    : num  14.17 13.95 8.98 17.32 11.5

If you are reading the data from a CSV file you should use something like:

table1 <- read.csv("my_data_file.csv",na.strings="*")

(header=TRUE is a default option for read.csv())

Notice that in the structure of the data, Air_Weight and Water_Weight are numeric (abbreviated num). This is good. We can proceed with a linear model:

fit <- lm(Water_Weight~Air_Weight, data=table1)

Answer 2

Try the following:

GENDER <- c(1,1,1,1,1)
Air_Weight <- c(75.60, 70.70, NA, 95.00, 73.20)
Water_Weight <- c(NA, 3.60, 4.00, 4.30, 3.80)
Body_Fat <- c(14.17, 13.95, 8.98, 17.32, 11.50)
ID <- c(01, 02, 03, 04, 05)
data <- data.frame(GENDER, Air_Weight, Water_Weight, Body_Fat)
data

This gives us the following:

       GENDER  Air_Weight   Water_Weight   Body_Fat
1      1       75.6          NA             14.17
2      1       70.7          3.6            13.95
3      1       NA            4.0            8.98
4      1       95.0          4.3            17.32
5      1       73.2          3.8            11.50

Then we fit it to the linear model with:

fit <- lm(Water_Weight~Air_Weight, data=data)
fit

And the output is:

lm(formula = Water_Weight ~ Air_Weight, data = data)

Coefficients:
(Intercept)   Air_Weight  
 1.7895       0.0265