yumyai / ksiga Goto Github PK

TODO:

We need to check for the correctness of the method, especially in the calculation part.

• Jensen-shannon distance
• Sparse jensen-shannon calculation
  • Is indices of final array still sorted?

See topic.

Add header about

• How/when output files are created
• Which version I use when running.

Objective:
User should be able to choose the distance function.

List:

1. Add option to choose a distance function to the program.
2. Change a default distance function to euclidian / cosine because they are fast.
3. Use sklearn.pairwise to calculate a distance.