blendtor
blendtor tries to detect the occurence of blend words by using approximate string matching algorithms.
Usage
In this example, we execute from a directory outside blendtor.py file.
python blendtor/blendtor.py [-m=mode] [-s=step]
m (mode): specific mode to run the program, for example jw (for Jaro-Winkler), ngram, and led (for Local Edit Distance)
s (step): parameter used to speed up execution by reducing the number of candidate words used. This value is 1, by default, processing all candidate values.
python blendtor/analysis.py
The above command will return string containing statistical properties to standard output. Statistical properties include central values (mean, median, mode) and dispersion/variability (variance and standard deviation).
Example
$ python blendtor/blendtor.py -m=jw -s=8000
precision 0.0 (0/3)
recall 0.0 (0/183)
$ python blendtor/analysis.py
statistics for JW:
mean : 0.7423989565383008
median : 0.7833333333333333
mode : 0.8333333333333334
stdev : 0.1729905416565893
pstdev : 0.17275405413432723
variance : 0.029925727502640165
pvariance : 0.029843963219846066
statistics for NGram:
mean : 6.8743169398907105
median : 6.0
mode : 4
stdev : 2.987746014051291
pstdev : 2.983661602006302
variance : 8.926626244479376
pvariance : 8.90223655528681
statistics for LED:
mean : 3.8224043715846996
median : 4.0
mode : 3.0
stdev : 1.7481546310660918
pstdev : 1.7457648081701527
variance : 3.0560446141178232
pvariance : 3.04769476544537