Conclusion

Opening of the libraries

We use the following libraries

library(car)
library(lmtest)

Dataset Import

rm(list=ls())
adot <- read.csv2("Life_Expectancy_Data.csv",header=TRUE,sep=",",dec=".")
head(adot)

NA

We want the data just from the year 2011 (Data Selection)

attach(adot)
adot <- adot[Year==2011,]
adot

We want to choose just some of the columns

adot <- adot[,c("Life_expectancy","Alcohol","percentage_expenditure","BMI")]

If we have any empty spaces in the database, than we delete the corresponding records

adot <- na.omit(adot)

Making some basic descriptive statistics

summary(adot)

 Life_expectancy    Alcohol       percentage_expenditure      BMI       
 Min.   :48.90   Min.   : 0.010   Min.   :    0.00       Min.   : 2.50  
 1st Qu.:64.30   1st Qu.: 1.230   1st Qu.:   20.83       1st Qu.:22.00  
 Median :73.40   Median : 4.090   Median :  158.28       Median :45.30  
 Mean   :70.78   Mean   : 4.887   Mean   : 1039.82       Mean   :39.85  
 3rd Qu.:76.10   3rd Qu.: 8.110   3rd Qu.:  683.92       3rd Qu.:57.90  
 Max.   :88.00   Max.   :17.310   Max.   :18822.87       Max.   :75.70  

Running the OLS (Ordinary Least Squares) regression

attach(adot)
#GDPPerCapita <- adot$GDP / adot$Polulation
result <-lm(Life_expectancy ~ +1 + Alcohol + BMI + percentage_expenditure )
summary(result)

Call:
lm(formula = Life_expectancy ~ +1 + Alcohol + BMI + percentage_expenditure)

Residuals:
     Min       1Q   Median       3Q      Max 
-17.2645  -4.0634   0.6862   4.0606  17.3444 

Coefficients:
                        Estimate Std. Error t value Pr(>|t|)    
(Intercept)            5.993e+01  1.190e+00  50.356  < 2e-16 ***
Alcohol                5.215e-01  1.377e-01   3.786 0.000209 ***
BMI                    1.879e-01  2.582e-02   7.277 1.07e-11 ***
percentage_expenditure 7.888e-04  2.057e-04   3.834 0.000175 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.736 on 177 degrees of freedom
Multiple R-squared:  0.4344,    Adjusted R-squared:  0.4248 
F-statistic: 45.31 on 3 and 177 DF,  p-value: < 2.2e-16

Based on the estimations, all of the regressors positivelly influnce the Life Expectancy (unfortunatelly, we are not able to explain this fact). It is recommended to include some other additional variables from the original database.

plot(result)

Interpret the graphs…