On the extension of fuzzy k-means algorithms for detection of linear clusters

@article{Bezdek1978OnTE,
  title={On the extension of fuzzy k-means algorithms for detection of linear clusters},
  author={James C. Bezdek and Robert W. Gunderson and Robert Ehrlich and Tom Meloy},
  journal={1978 IEEE Conference on Decision and Control including the 17th Symposium on Adaptive Processes},
  year={1978},
  pages={1438-1443}
}
In this note we discuss an extension of the fuzzy ISODATA (or fuzzy k-means) clustering algorithms, which attempt to recognize and compensate for their failure at detection of linear substructure in finite data sets. A numerical example is given to substantiate the proposed extension: and we offer some theoretical conjectures concerning its construction.