https://github.com/yhsunshining/fuzzy-c-means
FCM algorithm based on Tabu Search algorithm for fuzzy clustering
https://github.com/yhsunshining/fuzzy-c-means
fuzzy-cmeans tabu
Last synced: about 1 year ago
JSON representation
FCM algorithm based on Tabu Search algorithm for fuzzy clustering
- Host: GitHub
- URL: https://github.com/yhsunshining/fuzzy-c-means
- Owner: yhsunshining
- License: mit
- Created: 2017-03-07T01:42:51.000Z (over 9 years ago)
- Default Branch: master
- Last Pushed: 2018-05-01T12:43:52.000Z (about 8 years ago)
- Last Synced: 2024-10-03T12:30:42.502Z (over 1 year ago)
- Topics: fuzzy-cmeans, tabu
- Language: Python
- Homepage:
- Size: 1.86 MB
- Stars: 4
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.MD
- License: LICENSE
Awesome Lists containing this project
README
# fuzzy C-Means with Tabu Search
## current Problem
1. ~~How to generate the neighbourhood with the local solution?~~
> F: Use random to generate a mat, plus the mat and the solution. This dot is in a hyperrectangle neighbouthood of the mat.
2. ~~How to define the tabu object, code or the mat?
if use the mat, what should we define the object, the mat itself or the area ?~~
> F: use mat and tabu the area of a rect around the mat(see mat as a dot in N-d space)
3. Calc the J first or judge the tabu-list first?
4. When get the minimum J, the FMI isn't the maximum. When the FMI is maximum, the J isn't the minimum. What should we choose?
5. How to comprehend the saddle point in iteration seq?
6. ~~How to comprehend that the local minimun is related in the init value in iteration seq?~~
> F: In 3-d space ,z in the value of J, consider that there is a local minimun in the 2nd quadrant and a global minimun in the 4th quadrant. When y0>0, it's easy to find the local mininum and never find the global mininum.