This Pull Request adds to the repository a fix for the K-means clustering key generation algorithm that didn't work correctly. A short interval was interpreted as 0, a long interval as 1. Therefore long was encoded the same as short-short-short-long (both 1). To fix this there is a very large offset in the final encoding (10^10) at which position the number of intervals is placed.
Why
This Pull Request is needed because the algorithm didn't work correctly
How
This feature can be viewed/tested within the project by running the new tests
Alternative implementation
Other implementations that I've have considered are ...
Relevant Issues: -
Related Pull Requests: -
What
This Pull Request adds to the repository a fix for the K-means clustering key generation algorithm that didn't work correctly. A short interval was interpreted as 0, a long interval as 1. Therefore long was encoded the same as short-short-short-long (both 1). To fix this there is a very large offset in the final encoding (10^10) at which position the number of intervals is placed.
Why
This Pull Request is needed because the algorithm didn't work correctly
How
This feature can be viewed/tested within the project by running the new tests
Alternative implementation
Other implementations that I've have considered are ...
Notes
None