前女友tract Cluster Prototypes component does not show my id attribute

JeffersonjpaJeffersonjpa MemberPosts:5Contributor I
edited April 2019 inHelp
How could I pass an attribute with id label through an Extract Cluster Prototypes?
I need to identify the centers of the clusters (centroids) after the process of clustering with k-medoids for this my dataset has an identifier attribute that I set up as being id label using the setRole operator but the Extract Cluster Prototypes component does not show my id attribute. Can someone help me ?

Answers

  • varunm1varunm1 Moderator, MemberPosts:1,207Unicorn
    Hello@Jeffersonjpa

    对不起,我对这个问题有点困惑。前女友tract cluster prototypes will get the centroids of each attribute for different cluster independent of a label. This operator will have your label but it is used only for visualization purpose. If you want the labels and cluster ID you should connect the output (Clustering.clustered set) of k-medoids to result. This will show the attributes, cluster of each sample and label assigned.
    Regards,
    Varun
    https://www.varunmandalapu.com/

    Be Safe. Follow precautions and Maintain Social Distancing

    Jeffersonjpa
  • JeffersonjpaJeffersonjpa MemberPosts:5Contributor I
    edited April 2019
    Thank you@varunm1but I'm afraid this is not the answer to my problem..:D
    前女友plaining best, my dataset has already an identifier attribute I want to see this same unchanged attribute identifier at the end of the clustering process as a way to identify centroid (using k-medoids) within my dataset and complementing the problem I'm using normalization to improve the distance calculation between the attributes and my prior identifier should not be normalized
  • Telcontar120Telcontar120 Moderator, RapidMiner Certified Analyst, RapidMiner Certified Expert, MemberPosts:1,635Unicorn
    The centroid is designed to be an average of all your attributes for each cluster, so it does not output an id. There is only one row per cluster.
    If you have used k-medoids, then you can use Join to pull the averages into your full dataset and then map those centroids back to specific examples by Generate Attrbiutes.
    This will show you which individual records match your cluster centroid. See attached example process:
    < ?xml version = " 1.0 " encoding = " utf - 8 " ?> <过程版本sion="9.2.001">                                                                         


    Brian T.
    Lindon Ventures
    Data Science Consulting from Certified RapidMiner Experts
    varunm1
Sign InorRegisterto comment.