The who.csv dataset contains real-world data from 2008. The variables included follow.
Country: name of the country
LifeExp: average life expectancy for the country in years
InfantSurvival: proportion of those surviving to one year or more
Under5Survival: proportion of those surviving to five years or more
TBFree: proportion of the population without TB.
PropMD: proportion of the population who are MDs
PropRN: proportion of the population who are RNs
PersExp: mean personal expenditures on healthcare in US dollars at average exchange rate
GovtExp: mean government expenditures per capita on healthcare, US dollars at average exchange rate
TotExp: sum of personal and government expenditures.
Interpretation1
R_squared is considerable low (0.2577), which means the model can only explain 25% of data variability. The F statistic is not very representative as it is a one-factor model. The Std Error of the coefficient TotExp is about 8.5 times smaller than the coefficient which is good in terms of variability size of the slope. The p-value of the coefficient TotExp is very low, which means it is very relevant to the prediction of the target variable in the model.
# Residuals plot shows a very disperse variability with a clear pattern (not centered around zero). # Q-Q plot shows the residuals clearly deviating from the theoretical straight line
Interpretation2
R_squared went up considerably from (0.2577) to (0.7298), which means the model can only explain almost 73% of data variability. The F statistic is not very representative as it is a one-factor model. The Std Error of the coefficient TotExp is about 22 times smaller than the coefficient which is good in terms of variability size of the slope. The p-value of the coefficient TotExp is much smaller (2e-16), still very relevant to the prediction of the target variable in the model.
Residuals plot shows a relatively constant variability with no apparent patterns. Q-Q plot shows the residuals relatively following the theoretical straight line (except on the ends), which denotes a normal distribution
This model performs much better than the first degree one
3. Using the results from 3, forecast life expectancy when TotExp^.06 =1.5. Then forecast life expectancy when TotExp^.06=2.5.
test_data <- data.frame(TotExp_powered=c(1.5, 2.5))
predicted.dat <- predict(who_lm2, newdata=test_data, interval='confidence')^(1/4.6)
predicted.dat
## fit lwr upr
## 1 63.31153 62.10686 64.43893
## 2 86.50645 85.41583 87.54970
For TotExp = 1.5, LifeExp = 63.3 with a Confidence Interval (62.1, 64.4) at 95% Confidence level
For TotExp = 2.5, LifeExp = 86.5 with a Confidence Interval (85.4, 87.5) at 95% Confidence level
4. Build the following multiple regression model and interpret the F Statistics, R^2, standard error, and p-values. How good is the model?
LifeExp = b0+b1 x PropMd + b2 x TotExp +b3 x PropMD x TotExp
who_lm3 <- lm(LifeExp ~ PropMD + TotExp + PropMD*TotExp, data=who)
summary(who_lm3)
##
## Call:
## lm(formula = LifeExp ~ PropMD + TotExp + PropMD * TotExp, data = who)
##
## Residuals:
## Min 1Q Median 3Q Max
## -27.320 -4.132 2.098 6.540 13.074
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.277e+01 7.956e-01 78.899 < 2e-16 ***
## PropMD 1.497e+03 2.788e+02 5.371 2.32e-07 ***
## TotExp 7.233e-05 8.982e-06 8.053 9.39e-14 ***
## PropMD:TotExp -6.026e-03 1.472e-03 -4.093 6.35e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 8.765 on 186 degrees of freedom
## Multiple R-squared: 0.3574, Adjusted R-squared: 0.3471
## F-statistic: 34.49 on 3 and 186 DF, p-value: < 2.2e-16
# Residuals Analysis
plot(who_lm3$fitted.values, who_lm3$residuals, xlab='Fitted Values', ylab='Residuals')
abline(0,0, col="red")

qqnorm(who_lm3$residuals)
qqline(who_lm3$residuals)

Interpretation3
R_squared went down considerably from (0.7298) to (0.3574), which means the model can only explain almost 36% of data variability. The F statistic is 34.5 by considering 3 variables. The Std Error of the first 2 coefficients are several times smaller than the coefficients except for the last combined feature. The p-values of the coefficients very small so all variables are very relevant to the prediction of the target variable in the model.
# Residuals plot shows a very disperse variability with a clear pattern (not centered around zero). # Q-Q plot shows the residuals clearly deviating from the theoretical straight line
This model is not very good, very similar to the first model with 1 variable of order 1
5. Forecast LifeExp when PropMD=.03 and TotExp = 14. Does this forecast seem realistic? Why or why not?
test_data <- data.frame(PropMD=0.03, TotExp=14)
predicted.dat <- predict(who_lm3, newdata=test_data, interval='confidence')
predicted.dat
## fit lwr upr
## 1 107.696 91.85996 123.5321
For PropMD = 0.03 & TotExp = 14, LifeExp = 107.6 with a Confidence Interval (91.8, 123.5) at 95% Confidence level
The prediction and Confidence Interval seem unrealistically high, primarily because the LifeExp caps up to 83 and the TotExp of 14 is very low compared to the training set numbers (min of 13)