| URL: | https://github.com/zislam/GenClustPlusPlus/ | |
| Author: | Michael Furner | |
| Maintainer: | Michael Furner <mfurner{[at]}csu.edu.au> |
Implements clustering based on genetic algorithm for centroid generation based on Islam, M. Z., Estivill-Castro, V., Rahman, M. A. and Bossomaier, T. (2018). Combining K-Means and a Genetic Algorithm through a Novel Arrangement of Genetic Operators for High Quality Clustering. Expert Systems with Applications. Differences to the original algorithm are: 1. No use of VICUS similarity measure - standard EuclideanDistance class is used instead. 2. Uses basic missing value handling from SimpleKMeans. 3. If an operation generates a chromosome where all records are assigned to a single cluster, chromosome will be mutated until at least 2 clusters are found. 4. The starting generation for the chromosome selection operation is now modifiable, where in the original paper it was set to 11. The default is now after generation 50 with the default number of generations being 60.