r - Force a LM coefficient to a value -

this question has answer here:

• constrained linear regression coefficients in r [duplicate] 2 answers

i working on project weights screws.

i looped through , found fits number of combinations shown below.

head style       measurement  intercept     slope                1            x             1.0                2            x             2.2                3            x             4.1 b                1            x             1.2 b                2            x             2.0 b                3            x             4.1 

if know fact ones same measurement should have same slope. a-1 should have same slope b-1. there way run lm set slopes set number?

it's still not entirely clear you're asking, let's suppose original data looks this:

head_style measurement  x  y          1            .  .          1            .  .          1            .  .          2            .  . ... b          1            .  . 

so you've done far fit separate models (with y~x) each combination of measurement , head_style. suppose want slope depend on measurement , intercept depend on both measurement , head_slope.

first, make sure measurement factor

my_dat <- transform(my_dat,measurement=factor(measurement)) 

now fit single lm model:

lm(y~head_style:measurement + measurement:x, data=my_data) 

this should give separate intercept each combination of head_style , measurement , different slope each value of measurement.