[
https://issues.apache.org/jira/browse/FLINK-1745?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14901837#comment-14901837
]
Daniel Blazevski commented on FLINK-1745:
-----------------------------------------
Hi [~chiwanpark] and [~till.rohrmann],
I have implemented a Quadtree data structure to put into the KNN algorithm --
R-Tree version may follow. I will take each Block of the training DataSet in
the existing KNN and make a quadtree before finding the KNN on each block.
The quadtree class that I have is now a prototype -- using only
Arrays/ListBuffer[Double] types and not yet any Flink data structures -- but
before I transfer to Flink, let me know if you'd like the quadtree class put in
a specific directory -- otherwise I will just put in the in the same directory
as where KNN.scala is in (or maybe create a sub-directory where KNN.scala is)
Let me also generally know if there is anything specif you'd like to see done
to the quadtree when going from basic Scala code and importing to Flink;
otherwise I will look at other Flink code as a guide. I have made a simple
test of the quadtree, which I can also expand upon and include if you'd like.
Cheers,
Dan
> Add exact k-nearest-neighbours algorithm to machine learning library
> --------------------------------------------------------------------
>
> Key: FLINK-1745
> URL: https://issues.apache.org/jira/browse/FLINK-1745
> Project: Flink
> Issue Type: New Feature
> Components: Machine Learning Library
> Reporter: Till Rohrmann
> Assignee: Daniel Blazevski
> Labels: ML, Starter
>
> Even though the k-nearest-neighbours (kNN) [1,2] algorithm is quite trivial
> it is still used as a mean to classify data and to do regression. This issue
> focuses on the implementation of an exact kNN (H-BNLJ, H-BRJ) algorithm as
> proposed in [2].
> Could be a starter task.
> Resources:
> [1] [http://en.wikipedia.org/wiki/K-nearest_neighbors_algorithm]
> [2] [https://www.cs.utah.edu/~lifeifei/papers/mrknnj.pdf]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
