Can anybody explain how to assign cluster vectors to appropriate cluster centers after performing Kmeans clustering? I am working with R/Rstudio, is there any library that does it? or can anybody direct me to useful materials that could guide me? The essence is to get back my data from m by n(original data) to m by n-x (reduced data) after performing PCA and K-means.
Thanks
Dear Folasade Isinkaye,
Look the link, may be useful.
https://www.coursehero.com/file/p5bef3c/rather-than-randomly-assigning-all-the-vectors-to-clusters-at-the-start/
http://www.stat.cmu.edu/~cshalizi/350/lectures/07/lecture-07.pdf
Regards, Shafagat
Dear Folasade Isinkaye,
There are two solutions to identify the cluster centroid over the actual dataset.
Solution-1:
Trace the distinct patterns (Data Object) over the spatial dataset (Actual) in manually and desired the distinct number of data object could define as a cluster centroids.
Solution-2:
There is an existing schemes (Nearest Neighbour (NN), K-NN) are available to identify the distinct data object over the spatial dataset and you can follow to desired the appropriate number of data objects could act as cluster centroids.
You may use the package CARET for machine learning purpose. It has an extensive use for classification and clustering problems in R. Your input data should be also vectors.
https://docs.scipy.org/doc/scipy/reference/cluster.vq.html
- Badges
- Science topic
- Similar topics
- Books