Developing a C program to extract entries from census records to form a list of potential links in preparation for ML training.
To run the program, enter the src/ directory.
Copy template to a file named config and edit this file as needed.
Finally, execute ./run.sh.
This shell script reads the configuration file, sorts the input data
(if necessary), and executes the program.
You may want to redirect stdout and stderr to a file.
Using bash, you can do ./run.sh &> file.
OpenMP support is required for parallel performance.
To use OpenMP, please make sure you have gcc installed on your machine and not
a symlink to Clang. You can check this with gcc --version.
To install gcc, you will need Homebrew.
Install gcc using brew install gcc --without-multilib.
This should install gcc in /usr/local/bin as gcc-X, where X
is the version number.
Add /usr/local/bin to your path if it is not there already.
Finally, in src/Makefile, change the first line to read CC=gcc-X.
- Feigenbaum, James J. "Automated census record linking: A machine learning approach." (2016).