NBA-Search
NBA-Search copied to clipboard
skekre98
Implement optimal agglomerative clustering
Description of feature/enhancement This function should create the optimal clusters via agglomerative clustering in the ClusterStore.
Description of implementation
https://github.com/skekre98/NBA-Search/blob/0564a752c77a7c420890c44e6873c53111251fe5/inference/cluster_store.py#L78
There is likely an elegant way of doing this with sklearn without having to find the optimal number of clusters manually. See build_optimal_kmeans_clusters for how this was done for KMeans. This function should follow a Heirarchical Approach and start of treating each object as a singleton cluster.
Why this is a good feature/enhancement Would be nice to see how different clustering algorithms perform on different sets of data.
Additional context n/a
@skekre98 i can do this. Need some times (approximate 1 week). Currently I'm little bit busy with some other staff.
@RaihanSabique no worries, take you time ☺️. Thanks for seeing the entire component through!
Hi @skekre98, is this issue available? Would love to work on it but might need a couple weeks to get familiar with agglomerative clustering. @RaihanSabique, are you still trying to implement this?
Sorry. I was little busy with my graduate admission. Hope now i can help @skekre98
Sorry. I was little busy with my graduate admission. Hope now i can help @skekre98
Hi @RaihanSabique, please go ahead! I think @kelaaron is also interested in working on this if you'd like to guide him on what you've already done with the other functions.
