Regression on subset of data set

问题

I'd like to do the following and need some help:

Calculate slope and intercept for "Height" over "Age" [lm(Height~Age)] separately for

(A) each individual

(B) gender

and create a table containing the results (slope and intercept). Can I use "apply" for this?

In a next step I would like to do a statistical test to determine if slope and intercept are significantly different between Gender. I know how to do the test in R but maybe there is a way to combine slope/intercept calculation and T-testing.

Example data:

example = data.frame(Age = c(1, 3, 6, 9, 12,
                             1, 3, 6, 9, 12,
                             1, 3, 6, 9, 12,
                             1, 3, 6, 9, 12), 
                Individual = c("Jack", "Jack", "Jack", "Jack", "Jack",
                               "Jill", "Jill", "Jill", "Jill", "Jill",
                               "Tony", "Tony", "Tony", "Tony", "Tony",
                               "Jen", "Jen", "Jen", "Jen","Jen"),
                    Gender = c("M", "M", "M", "M", "M",
                               "F", "F", "F", "F", "F",
                               "M", "M", "M", "M", "M",
                               "F", "F", "F", "F", "F"),
                    Height = c(38, 62, 92, 119, 165,
                               31, 59, 87, 118, 170,
                               45, 72, 93, 155, 171,
                               33, 61, 92, 115, 168))

回答1:

One way to do regression analysis separately for each level and then combine slopes and intercepts in data frame, is to use function ddply() from library plyr.

library(plyr)

ddply(example,"Individual",function(x) coefficients(lm(Height~Age,x)))
  Individual (Intercept)      Age
1       Jack    26.29188 11.11421
2        Jen    22.10660 11.56345
3       Jill    18.33249 12.04315
4       Tony    33.02030 11.96447

ddply(example,"Gender",function(x) coefficients(lm(Height~Age,x)))
  Gender (Intercept)      Age
1      F    20.21954 11.80330
2      M    29.65609 11.53934

来源：https://stackoverflow.com/questions/14471351/regression-on-subset-of-data-set

标签

linear-regression