CODEWITHSUNDEEP
rlinear-regression
user7220615
user7220615

Reputation:

Prediction with regression equation in R

First of all let me tell you I'm very new to R, so my question might sounds weird to you, I want to predict the weight of an individual from the height. For that I need to find the equation of the regression line (in the following form) for this problem:

Weight = intercept + (slope) x Height

I've the following form of data ?

Any idea about how to find regression line equation

dat <- read.table(text = "SampleNo,Height,Weight
                          1,65.78,112.99
                          2,71.52,136.49
                          3,69.40,153.03
                          4,68.22,142.34
                          5,67.79,144.30
                          6,68.70,123.30
                          7,69.80,141.49
                          8,70.01,136.46
                          9,67.90,112.37", 
                  sep = ",", header = T)

Upvotes: 1

Views: 1285

Answers (1)

Nick Criswell
Nick Criswell

Reputation: 1743

You can accomplish this with the lm function. 

lm1 <- lm(Weight ~ Height, data = dat )

I want to predict Weight as a function of Height so I use the syntax Weight ~ Height.

Finally, I run the coefficients function on the lm1 object to get the coefficient for Height and the intercept. 

coefficients(lm1)
    (Intercept)      Height 
    -177.650244    4.525168

If I want to predict the result for Weight for a certain set of Heights, I can do that with the following:

> predict(lm1, newdata = data.frame((Height = c(65, 68.5, 71.6))))
       1        2        3 
116.4857 132.3238 146.3518

You can get additional information about the fit using the summary function on the lm1 object. This will provide R^2 values, more info on the estimates, etc...

summary(lm1)

Call:
lm(formula = Weight ~ Height, data = vals)

Residuals:
    Min      1Q  Median      3Q     Max 
-17.239  -9.500  -2.697  11.283  16.634 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) -177.650    193.247  -0.919    0.389
Height         4.525      2.808   1.611    0.151

Residual standard error: 13.03 on 7 degrees of freedom
Multiple R-squared:  0.2705,    Adjusted R-squared:  0.1663 
F-statistic: 2.596 on 1 and 7 DF,  p-value: 0.1512

A great (free) resource for regression and many other machine learning techniques in R can be found here. It's a great book to learn about stats and how to implement those methods in R.

Upvotes: 1

Related Questions