R's standard way of doing regression on categorical variables is to select one factor level as a reference level and constraining the effect of that level to be zero. Instead of constraining a single level effect to be zero, I'd like to constrain the sum of the coefficients to be zero.
I can hack together coefficient estimates for this manually after fitting the model the standard way:
x <- lm(data = mtcars, mpg ~ factor(cyl))
z <- c(coef(x), "factor(cyl)4" = 0)
y <- mean(z[-1])
z[-1] <- z[-1] - y
z[1] <- z[1] + y
z
## (Intercept) factor(cyl)6 factor(cyl)8 factor(cyl)4
## 20.5021645 -0.7593074 -5.4021645 6.1614719
But that leaves me without standard error estimates for the former reference level that I just added as an explicit effect, and I need to have those as well.
I did some searching and found the constrasts functions, and tried
lm(data = mtcars, mpg ~ C(factor(cyl), contr = contr.sum))
but this still only produces two effect estimates. Is there a way to change which constraint R uses for linear regression on categorical variables properly?
Think I've figured it out. Using contrasts actually is the right way to go about it, you just need to do a little work to get the results into a convenient looking form. Here's the fit:
fit <- lm(data = mtcars, mpg ~ C(factor(cyl), contr = contr.sum))
Then the matrix cs <- contr.sum(factor(cyl)) is used to get the effect estimates and the standard error.
The effect estimates just come from multiplying the contrast matrix by the effect estimates lm spits out, like so:
cs %*% coef(fit)[-1]
The standard error can be calculated using the contrast matrix and the variance-covariance matrix of the coefficients, like so:
diag(cs %*% vcov(fit)[-1,-1] %*% t(cs))
Related
I modelled a multiple Regression based on the Mincer-Wage-Equation and I added a weighting-factor to make it representative for the whole population.
But when I'm adding the weights function into my modell, R calculates less standardized residuals than I have observations.
Here's my modell:
lm(log(earings) ~ Gender + Age + Age^2 + Education, weights= phrf)
So I got problems to analyze the residuals because when I'm trying to plot the rstandard against the fitted.values R is telling: Different Variable Length in rstandard() found.
This Problem ist only by rstandard and rstudent, when I'm plotting the normal resid() against fitted.values there is no problem.
And when I'm leaving out the weights function I have not problems, too.
In the help file for rstudent():
Note that cases with weights == 0 are dropped from all these functions, but that if a linear model has been fitted with na.action = na.exclude, suitable values are filled in for the cases excluded during fitting.
A simple example to demonstrate:
set.seed(123)
x <- 1:100
y <- x + rnorm(100)
w <- runif(100)
w[44] <- 0
fit <- lm(y ~ x, weights=w)
length(fitted(fit))
length(rstudent(fit))
Gives:
> length(fitted(fit))
[1] 100
> length(rstudent(fit))
[1] 99
And this makes sense. If you have a weight of 0, the theoretical variance is 0 which is an infinite studentized or standardized residual.
Since you are effectively deleting those observations, you can subset the call to lm with subset=w!=0 or you can use that flag for the fitted values:
plot(fitted(fit)[w!=0], rstudent(fit))
Is there a way how I can extract coefficients of globally fitted terms in local regression modeling?
Maybe I do misunderstand the role of globally fitted terms in the function loess, but what I would like to have is the following:
# baseline:
x <- sin(seq(0.2,0.6,length.out=100)*pi)
# noise:
x_noise <- rnorm(length(x),0,0.1)
# known structure:
x_1 <- sin(seq(5,20,length.out=100))
# signal:
y <- x + x_1*0.25 + x_noise
# fit loess model:
x_seq <- seq_along(x)
mod <- loess(y ~ x_seq + x_1,parametric="x_1")
The fit is done perfectly, however, how can I extract the estimated value of the globally fitted term x_1 (i.e. some value near 0.25 for the example above)?
Finally, I found a solution to my problem using the function gam from the package gam:
require(gam)
mod2 <- gam(y ~ lo(x_seq,span=0.75,degree=2) + x_1)
However, the fits from the two models are not exactly the same (which might be due to different control settings?)...
I fitted a model in R with the lmer()-function from the lme4 package. I scaled the dependent variable:
mod <- lmer(scale(Y)
~ X
+ (X | Z),
data = df,
REML = FALSE)
I look at the fixed-effect coefficients with fixef(mod):
> fixef(mod)
(Intercept) X1 X2 X3 X4
0.08577525 -0.16450047 -0.15040043 -0.25380073 0.02350007
It is quite easy to calculate the means by hand from the fixed-effects coefficients. However, I want them to be unscaled and I am unsure how to do this exactly. I am aware that scaling means substracting the mean from every Y and deviding by the standard deviation. But both, mean and standard deviation, were calculated from the original data. Can I simply reverse this process after I fitted an lmer()-model by using the mean and standard deviation of the original data?
Thanks for any help!
Update: The way I presented the model above seems to imply that the dependent variable is scaled by taking the mean over all responses and dividing by the standard deviation of all the responses. Usually, it is done differently. Rather than taking the overall mean and standard deviation the responses are standardized per subject by using the mean and standard deviation of the responses of that subject. (This is odd in an lmer() I think as the random intercept should take care of that... Not to mention the fact that we are talking about calculating means on an ordinal scale...) The problem however stays the same: Once I fitted such a model, is there a clean way to rescale the coefficients of the fitted model?
Updated: generalized to allow for scaling of the response as well as the predictors.
Here's a fairly crude implementation.
If our original (unscaled) regression is
Y = b0 + b1*x1 + b2*x2 ...
Then our scaled regression is
(Y0-mu0)/s0 = b0' + (b1'*(1/s1*(x1-mu1))) + b2'*(1/s2*(x2-mu2))+ ...
This is equivalent to
Y0 = mu0 + s0((b0'-b1'/s1*mu1-b2'/s2*mu2 + ...) + b1'/s1*x1 + b2'/s2*x2 + ...)
So bi = s0*bi'/si for i>0 and
b0 = s0*b0'+mu0-sum(bi*mui)
Implement this:
rescale.coefs <- function(beta,mu,sigma) {
beta2 <- beta ## inherit names etc.
beta2[-1] <- sigma[1]*beta[-1]/sigma[-1]
beta2[1] <- sigma[1]*beta[1]+mu[1]-sum(beta2[-1]*mu[-1])
beta2
}
Try it out for a linear model:
m1 <- lm(Illiteracy~.,as.data.frame(state.x77))
b1 <- coef(m1)
Make a scaled version of the data:
ss <- scale(state.x77)
Scaled coefficients:
m1S <- update(m1,data=as.data.frame(ss))
b1S <- coef(m1S)
Now try out rescaling:
icol <- which(colnames(state.x77)=="Illiteracy")
p.order <- c(icol,(1:ncol(state.x77))[-icol])
m <- colMeans(state.x77)[p.order]
s <- apply(state.x77,2,sd)[p.order]
all.equal(b1,rescale.coefs(b1S,m,s)) ## TRUE
This assumes that both the response and the predictors are scaled.
If you scale only the response and not the predictors, then you should submit (c(mean(response),rep(0,...)) for m and c(sd(response),rep(1,...)) for s (i.e., m and s are the values by which the variables were shifted and scaled).
If you scale only the predictors and not the response, then submit c(0,mean(predictors)) for m and c(1,sd(predictors)) for s.
Let me state my confusion with the help of an example,
#making datasets
x1<-iris[,1]
x2<-iris[,2]
x3<-iris[,3]
x4<-iris[,4]
dat<-data.frame(x1,x2,x3)
dat2<-dat[1:120,]
dat3<-dat[121:150,]
#Using a linear model to fit x4 using x1, x2 and x3 where training set is first 120 obs.
model<-lm(x4[1:120]~x1[1:120]+x2[1:120]+x3[1:120])
#Usig the coefficients' value from summary(model), prediction is done for next 30 obs.
-.17947-.18538*x1[121:150]+.18243*x2[121:150]+.49998*x3[121:150]
#Same prediction is done using the function "predict"
predict(model,dat3)
My confusion is: the two outcomes of predicting the last 30 values differ, may be to a little extent, but they do differ. Whys is it so? should not they be exactly same?
The difference is really small, and I think is just due to the accuracy of the coefficients you are using (e.g. the real value of the intercept is -0.17947075338464965610... not simply -.17947).
In fact, if you take the coefficients value and apply the formula, the result is equal to predict:
intercept <- model$coefficients[1]
x1Coeff <- model$coefficients[2]
x2Coeff <- model$coefficients[3]
x3Coeff <- model$coefficients[4]
intercept + x1Coeff*x1[121:150] + x2Coeff*x2[121:150] + x3Coeff*x3[121:150]
You can clean your code a bit. To create your training and test datasets you can use the following code:
# create training and test datasets
train.df <- iris[1:120, 1:4]
test.df <- iris[-(1:120), 1:4]
# fit a linear model to predict Petal.Width using all predictors
fit <- lm(Petal.Width ~ ., data = train.df)
summary(fit)
# predict Petal.Width in test test using the linear model
predictions <- predict(fit, test.df)
# create a function mse() to calculate the Mean Squared Error
mse <- function(predictions, obs) {
sum((obs - predictions) ^ 2) / length(predictions)
}
# measure the quality of fit
mse(predictions, test.df$Petal.Width)
The reason why your predictions differ is because the function predict() is using all decimal points whereas on your "manual" calculations you are using only five decimal points. The summary() function doesn't display the complete value of your coefficients but approximate the to five decimal points to make the output more readable.
Suppose I have x values, y values, and expected y values f (from some nonlinear best fit curve).
How can I compute R^2 in R? Note that this function is not a linear model, but a nonlinear least squares (nls) fit, so not an lm fit.
You just use the lm function to fit a linear model:
x = runif(100)
y = runif(100)
spam = summary(lm(x~y))
> spam$r.squared
[1] 0.0008532386
Note that the r squared is not defined for non-linear models, or at least very tricky, quote from R-help:
There is a good reason that an nls model fit in R does not provide
r-squared - r-squared doesn't make sense for a general nls model.
One way of thinking of r-squared is as a comparison of the residual
sum of squares for the fitted model to the residual sum of squares for
a trivial model that consists of a constant only. You cannot
guarantee that this is a comparison of nested models when dealing with
an nls model. If the models aren't nested this comparison is not
terribly meaningful.
So the answer is that you probably don't want to do this in the first
place.
If you want peer-reviewed evidence, see this article for example; it's not that you can't compute the R^2 value, it's just that it may not mean the same thing/have the same desirable properties as in the linear-model case.
Sounds like f are your predicted values. So the distance from them to the actual values devided by n * variance of y
so something like
1-sum((y-f)^2)/(length(y)*var(y))
should give you a quasi rsquared value, so long as your model is reasonably close to a linear model and n is pretty big.
As a direct answer to the question asked (rather than argue that R2/pseudo R2 aren't useful) the nagelkerke function in the rcompanion package will report various pseudo R2 values for nonlinear least square (nls) models as proposed by McFadden, Cox and Snell, and Nagelkerke, e.g.
require(nls)
data(BrendonSmall)
quadplat = function(x, a, b, clx) {
ifelse(x < clx, a + b * x + (-0.5*b/clx) * x * x,
a + b * clx + (-0.5*b/clx) * clx * clx)}
model = nls(Sodium ~ quadplat(Calories, a, b, clx),
data = BrendonSmall,
start = list(a = 519,
b = 0.359,
clx = 2304))
nullfunct = function(x, m){m}
null.model = nls(Sodium ~ nullfunct(Calories, m),
data = BrendonSmall,
start = list(m = 1346))
nagelkerke(model, null=null.model)
The soilphysics package also reports Efron's pseudo R2 and adjusted pseudo R2 value for nls models as 1 - RSS/TSS:
pred <- predict(model)
n <- length(pred)
res <- resid(model)
w <- weights(model)
if (is.null(w)) w <- rep(1, n)
rss <- sum(w * res ^ 2)
resp <- pred + res
center <- weighted.mean(resp, w)
r.df <- summary(model)$df[2]
int.df <- 1
tss <- sum(w * (resp - center)^2)
r.sq <- 1 - rss/tss
adj.r.sq <- 1 - (1 - r.sq) * (n - int.df) / r.df
out <- list(pseudo.R.squared = r.sq,
adj.R.squared = adj.r.sq)
which is also the pseudo R2 as calculated by the accuracy function in the rcompanion package. Basically, this R2 measures how much better your fit becomes compared to if you would just draw a flat horizontal line through them. This can make sense for nls models if your null model is one that allows for an intercept only model. Also for particular other nonlinear models it can make sense. E.g. for a scam model that uses stricly increasing splines (bs="mpi" in the spline term), the fitted model for the worst possible scenario (e.g. where your data was strictly decreasing) would be a flat line, and hence would result in an R2 of zero. Adjusted R2 then also penalize models with higher nrs of fitted parameters. Using the adjusted R2 value would already address a lot of the criticisms of the paper linked above, http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2892436/ (besides if one swears by using information criteria to do model selection the question becomes which one to use - AIC, BIC, EBIC, AICc, QIC, etc).
Just using
r.sq <- max(cor(y,yfitted),0)^2
adj.r.sq <- 1 - (1 - r.sq) * (n - int.df) / r.df
I think would also make sense if you have normal Gaussian errors - i.e. the correlation between the observed and fitted y (clipped at zero, so that a negative relationship would imply zero predictive power) squared, and then adjusted for the nr of fitted parameters in the adjusted version. If y and yfitted go in the same direction this would be the R2 and adjusted R2 value as reported for a regular linear model. To me this would make perfect sense at least, so I don't agree with outright rejecting the usefulness of pseudo R2 values for nls models as the answer above seems to imply.
For non-normal error structures (e.g. if you were using a GAM with non-normal errors) the McFadden pseudo R2 is defined analogously as
1-residual deviance/null deviance
See here and here for some useful discussion.
Another quasi-R-squared for non-linear models is to square the correlation between the actual y-values and the predicted y-values. For linear models this is the regular R-squared.
As an alternative to this problem I used at several times the following procedure:
compute a fit on data with the nls function
using the resulting model make predictions
Trace (plot...) the data against the values predicted by the model (if the model is good, points should be near the bissectrix).
Compute the R2 of the linear régression.
Best wishes to all. Patrick.
With the modelr package
modelr::rsquare(nls_model, data)
nls_model <- nls(mpg ~ a / wt + b, data = mtcars, start = list(a = 40, b = 4))
modelr::rsquare(nls_model, mtcars)
# 0.794
This gives essentially the same result as the longer way described by Tom from the rcompanion resource.
Longer way with nagelkerke function
nullfunct <- function(x, m){m}
null_model <- nls(mpg ~ nullfunct(wt, m),
data = mtcars,
start = list(m = mean(mtcars$mpg)))
nagelkerke(nls_model, null_model)[2]
# 0.794 or 0.796
Lastly, using predicted values
lm(mpg ~ predict(nls_model), data = mtcars) %>% broom::glance()
# 0.795
Like they say, it's only an approximation.