Negative binomial regression fitting code in C++. Adapted from R source code.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
README.txt
daxpy.f
dcopy.f
ddot.f
dnrm2.f
dqrdc.f
dqrdc2.f
dqrls.f
dqrsl.f
dscal.f
dswap.f
fit.cpp
glm.nb.cpp
glm.nb.h
glm.nb.qsub.sh
lmfit.cpp
makefile

README.txt

Usage:
glm.nb [response file] [predictor file] [initial theta]

Inputs:
[response file]: A text file containing the responses (y vector) for each observation, one per line.

[predictor file]: A text file containing the feature values (x matrix) to use in the negative binomial fitting, where rows are observations, and columns are features.

[initial theta]: Used to determine convergence of the underlying algorithms. In R, this is typically set to 0.000001

Output:
"glm.nb" prints to standard output, which can be redirected. This output includes:
- Coefficients
- Residuals
- Fitted Values
- Effects
- R (correlation measure)
- Rank
- QR matrix
- Qraux
- Pivot vector
- Tol
- Linear predictors
- Deviance
- AIC
- Null deviance
- Number of iterations
- Weights
- Prior weights
- Degrees of freedom residual
- Degrees of freedom null
- Converged
- Boundary
- Theta
- SE Theta
- Two Log Likelihood