The objective of the program is to automatically cluster similar elements.
- The user can choose one of the files - "noraml.txt" or "unbalance.txt"
- The user can choose the count of the clusters.
- The program should use random restart, keeping the best solution in the new generations.
- The best solution is chosen based on internal cluster distance.
The program is using the algorithm K-means to automatically cluster similar elements with two attributes, provided as points in the Euclidean space in the files "normal.txt" and "unbalance.txt".
- "normal.txt" - 4 Gaussian clusters
- "unbalance.txt" - 8 Gaussian clusters
- Run
python3 main.py
- Input file name ("normal" or "unbalance")
- Input number of clusters
- The internal cluster distance of each generation is displayed.
- Image that shows the different clusters in different colors.
- Compare results with K-means++
- Compare results with Soft k-means
- Use intercluster distance.
- Use combination of both internal cluster distance and intercluster distance.