Scribd
Upload
41 views27 pages

Chapter3 Is This

The document discusses different types of intervals that can be calculated for linear regression models in R, including confidence and prediction intervals. It shows how to calculate confidence intervals for slope and intercept parameters using the tidy function. Bootstrap intervals for slopes are also demonstrated. Prediction intervals for the average response at given values of the predictor are calculated and plotted along with the linear fit and observed data points.

Uploaded by

Piel Survey
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
41 views27 pages

Chapter3 Is This

The document discusses different types of intervals that can be calculated for linear regression models in R, including confidence and prediction intervals. It shows how to calculate confidence intervals for slope and intercept parameters using the tidy function. Bootstrap intervals for slopes are also demonstrated. Prediction intervals for the average response at given values of the predictor are calculated and plotted along with the linear fit and observed data points.

Uploaded by

Piel Survey
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 27

Mathematical

approximation
INFERENCE FOR LINEAR REGRESSION IN R

Jo Hardin
Professor, Pomona College
INFERENCE FOR LINEAR REGRESSION IN R
Sampling distribution of slope: good t fit
ggplot(starFatCal, aes(x = statistic)) +
geom_histogram(aes(y = ..density..), bins = 50) +
stat_function(fun = dt, color = "red", args = list(df = nrow(starbucks) - 2))

INFERENCE FOR LINEAR REGRESSION IN R


INFERENCE FOR LINEAR REGRESSION IN R
Model fit

INFERENCE FOR LINEAR REGRESSION IN R


INFERENCE FOR LINEAR REGRESSION IN R
Sampling distribution of slope: poor t fit
ggplot(starProFib, aes(x = statistic)) + geom_histogram(aes(y = ..density..), bins = 50) +
stat_function(fun = dt, color = "red", args = list(df = nrow(starbucks) - 2))

INFERENCE FOR LINEAR REGRESSION IN R


Sampling distribution of slope: poor t fit

INFERENCE FOR LINEAR REGRESSION IN R


Sampling distribution of slope: poor t fit

INFERENCE FOR LINEAR REGRESSION IN R


Let's practice!
INFERENCE FOR LINEAR REGRESSION IN R
Intervals in
regression
INFERENCE FOR LINEAR REGRESSION IN R

Jo Hardin
Professor, Pomona College
INFERENCE FOR LINEAR REGRESSION IN R
INFERENCE FOR LINEAR REGRESSION IN R
alpha = 0.05
crit_val <- qt((1-alpha/2), df = nrow(starbucks) - 2)
lm(Calories ~ Fat, data=starbucks) %>%
tidy(conf.int = TRUE, conf.level = 1-alpha)

term estimate std.error statistic p.value conf.low conf.high


1 (Intercept) 147.9833 14.9719851 9.884013 6.630009e-17 118.31530 177.65128
2 Fat 12.7586 0.8171655 15.613236 8.937367e-30 11.13933 14.37787

lm(Calories ~ Fat, data=starbucks) %>% tidy() %>%


mutate(lower = estimate - crit_val*std.error,
upper = estimate + crit_val*std.error)

term estimate std.error statistic p.value lower upper


1 (Intercept) 147.9833 14.9719851 9.884013 6.630009e-17 118.31530 177.65128
2 Fat 12.7586 0.8171655 15.613236 8.937367e-30 11.13933 14.37787

INFERENCE FOR LINEAR REGRESSION IN R


Confidence interval for intercept parameter
tidy_mod <- lm(Calories ~ Fat, tidy_mod %>%
data = starbucks) %>% filter(term == "(Intercept)") %>%
tidy(conf.int = TRUE, select(conf.low, conf.high)
conf.level = 1-alpha)
tidy_mod
conf.low conf.high
1 118.3153 177.6513
term estimate std.error
1 (Intercept) 147.9833 14.9719851
2 Fat 12.7586 0.8171655
statistic p.value conf.low
1 9.884013 6.630009e-17 118.31530
2 15.613236 8.937367e-30 11.13933
conf.high
1 177.65128
2 14.37787

INFERENCE FOR LINEAR REGRESSION IN R


Confidence interval for slope parameter
tidy_mod <- lm(Calories ~ Fat, tidy_mod %>%
data = starbucks) %>% filter(term == "Fat") %>%
tidy(conf.int = TRUE, select(conf.low, conf.high)
conf.level = 1-alpha)
tidy_mod
conf.low conf.high
1 11.13933 14.37787
term estimate std.error
1 (Intercept) 147.9833 14.9719851
2 Fat 12.7586 0.8171655
statistic p.value conf.low
1 9.884013 6.630009e-17 118.31530
2 15.613236 8.937367e-30 11.13933
conf.high
1 177.65128
2 14.37787

INFERENCE FOR LINEAR REGRESSION IN R


Bootstrap interval for slope
BS_slope <- starbucks %>%
specify(Calories ~ Fat) %>%
generate(reps = 1000, type = "bootstrap") %>%
calculate(stat = "slope")
BS_slope %>%
summarize(low = quantile(stat, alpha / 2),
high = quantile(stat, 1 - alpha / 2))

A tibble: 1 x 2
low high
<dbl> <dbl>
1 11.16712 14.34817

INFERENCE FOR LINEAR REGRESSION IN R


Let's practice!
INFERENCE FOR LINEAR REGRESSION IN R
Different types of
intervals
INFERENCE FOR LINEAR REGRESSION IN R

Jo Hardin
Professor, Pomona College
INFERENCE FOR LINEAR REGRESSION IN R
INFERENCE FOR LINEAR REGRESSION IN R
Predicting average calories at specific fat
library(broom)
alpha <- .05
crit_val <- qt((1-alpha/2), df = nrow(starbucks) - 2)
newfood <- data.frame(Fat = c(0,10,20,30))
augment(lm(Calories ~ Fat, data=starbucks), newdata = newfood) %>%
mutate(lowMean = .fitted - crit_val*.se.fit,
upMean = .fitted + crit_val*.se.fit)

Fat .fitted .se.fit lowMean upMean


1 0 147.9833 14.971985 118.3153 177.6513
2 10 275.5693 8.516206 258.6938 292.4447
3 20 403.1552 7.378555 388.5341 417.7763
4 30 530.7412 13.035040 504.9114 556.5710

INFERENCE FOR LINEAR REGRESSION IN R


Creating CI for average response
predMeans <- augment(lm(Calories ~ Fat, data = starbucks)) %>%
select(Calories, Fat, .fitted, .se.fit) %>%
mutate(lowMean = .fitted - crit_val*.se.fit,
upMean = .fitted + crit_val*.se.fit)
head(predMeans)

Calories Fat .fitted .se.fit lowMean upMean


1 300 5 211.7763 11.473843 189.0401 234.5125
2 380 6 224.5349 10.823741 203.0869 245.9828
3 410 22 428.6724 8.176354 412.4704 444.8744
4 460 23 441.4310 8.663769 424.2632 458.5989
5 420 22 428.6724 8.176354 412.4704 444.8744
6 380 16 352.1209 6.756473 338.7324 365.5093

INFERENCE FOR LINEAR REGRESSION IN R


Plotting CI for average response
ggplot(predMeans, aes(x = Fat, y = Calories)) +
geom_point() +
stat_smooth(method = "lm", se = FALSE) +
geom_ribbon(aes(ymin = lowMean, ymax = upMean), alpha=.2)

INFERENCE FOR LINEAR REGRESSION IN R


alpha <- .05
crit_val <- qt((1-alpha/2), df = nrow(twins) - 2)
FatCal_lm <- lm(Calories ~ Fat, data = starbucks)
FatCal_gl <- glance(FatCal_lm)
FatCal_sig <- pull(FatCal_gl, sigma)
FatCal_pred <- augment(FatCal_lm) %>%
mutate(.se.pred = sqrt(FatCal_sig^2 + .se.fit^2))
predResp <- FatCal_pred %>%
mutate(lowResp = .fitted - crit_val*.se.pred, upResp = .fitted + crit_val*.se.pred)
predResp

A tibble: 113 x 12
Calories Fat .fitted .se.fit .resid .hat .sigma
<int> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 300 5 211.7763 11.473843 88.223722 0.025567957 71.57640
2 380 6 224.5349 10.823741 155.465125 0.022752704 70.50502
3 410 22 428.6724 8.176354 -18.672436 0.012983674 72.05959
... with 103 more rows, and 5 more variables: .cooksd <dbl>, .std.resid <dbl>, .se.pred <dbl>,
lowResp <dbl>, upResp bl>

INFERENCE FOR LINEAR REGRESSION IN R


Plotting prediction intervals
ggplot(predResp, aes(x = Fat, y = Calories)) +
geom_point() +
stat_smooth(method = "lm", se = FALSE) +
geom_ribbon(aes(ymin = lowResp, ymax = upResp), alpha = .2)

INFERENCE FOR LINEAR REGRESSION IN R


Let's practice!
INFERENCE FOR LINEAR REGRESSION IN R

You might also like