EECS 837 - Final project
Theodore Lindsey
About
This project reads lers-type files within the directory containing the python code and attempts to perform one of several discritizations on it:
- Global equal interval width
- Global equal frequency per interval
- Global conditional entropy
Requirements
This program is compatible with 3.x (a small amount of tweaking should make it compatible with python
2.x). It is necessary to have both
main.py
and
utility.py
in the same directory. You should put any lers-type files in the same directory
as the two code files. From there, simply run main.py and it should find
all .txt and .lers files. It also has an option for manual filename
entry for files in the same directory as the .py files.
Running
Use $ python3 ./main.py to get started.
Assignment
Notes
To the best of my knowledge, everything functions. Final remarks:
- equal frequency - The algorithm is wrong, I think. It has problems with attribute values that have a huge covering compared to the other attribute values and it skews results to the first groupings when this occurs
- python versions - It shouldn't be difficult to get this working with python2.x. In testing, the only problem appeared to be array indexes weren't being cast to ints properly. I suspect this is because python 3.x and 2.x handle division differently (/ vs //)