I have set up a MapReduce code for KMeans algorithm using python. The code is working fine.The resultant cluster file that is generated as part-00000 output file contains the set of data points with full dimension.Now I want to find out the index of the data points.How can I do that?
What do you mean by index of clusters?
KMeans cluster centers are not the points from the dataset.
You can find the index of cluster center which is closest to your test point (may be from train or test set).
As Dinesh said, it is true that the question is not clear enough. However I understand by that, that your are able to differentiate the data Points of each Cluster but unable to pick up each data point from the original data set and tell its Cluster number/associated centroid. For sequential as well as for parallel K-means algorithms you should always bind each data point with ist current Cluster number and original index. Since data points are usually of big dimensions, avoid moving them anyhow . Try to work with the original index of data Points. They are light and the Approach is flexible.
I don't know whether you are ready function or you write your own code for Kmeans. If you are using own code you can get data point index for each cluster. (Points which are closest center of cluster)
By the way if you want Kmeans code for Fortran Language, i can send you..
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Algorithms