tag:blogger.com,1999:blog-7583720.post9010037217765136815..comments2024-03-05T03:17:02.289-08:00Comments on Salmon Run: An implementation of the Silhouette Score metric on SparkSujit Palhttp://www.blogger.com/profile/06835223352394332155noreply@blogger.comBlogger8125tag:blogger.com,1999:blog-7583720.post-69822651527775802522021-02-27T11:46:24.307-08:002021-02-27T11:46:24.307-08:00Thanks, this was useful. Cleared up a few things f...Thanks, this was useful. Cleared up a few things for me.<br /><br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7583720.post-53959494267294387882020-04-29T11:56:00.024-07:002020-04-29T11:56:00.024-07:00Awesome news, thanks oskarryn! I will use that one...Awesome news, thanks oskarryn! I will use that one instead.Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-75726261774130324892019-06-12T03:34:43.538-07:002019-06-12T03:34:43.538-07:00Hi, just for info:
ClusteringEvaluator that comes ...Hi, just for info:<br />ClusteringEvaluator that comes with Spark 2.3 (released more or less when you published the post) has a scalable Silhouette implementation (https://issues.apache.org/jira/browse/SPARK-14516)<br /><br />oskarrynnoreply@blogger.comtag:blogger.com,1999:blog-7583720.post-27411525418799374412019-04-20T15:20:43.757-07:002019-04-20T15:20:43.757-07:00Here you have a PySpark implementation for the Sim...Here you have a PySpark implementation for the Simplified Silhouette Score.<br /><br />I started from the data arrangement proposed for Scala.<br /><br />Hope this helps.Anonymoushttps://www.blogger.com/profile/03632644497257052968noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-19395767049980108302018-03-14T13:04:52.905-07:002018-03-14T13:04:52.905-07:00Thanks, and no, I don't have a Python version,...Thanks, and no, I don't have a Python version, sorry. Although should be fairly easy to write (maybe easier because numpy has a cleaner API vs breeze IMO). I used Scala here mainly because you can inline code inside one of Spark's higher order functions, with Python you would need to write a separate function to compute the block that begins on line 40 and call it within the map call.Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-35297500358361676922018-03-13T19:52:47.915-07:002018-03-13T19:52:47.915-07:00Thanks so much for this! Do you have a Python vers...Thanks so much for this! Do you have a Python version of this?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7583720.post-31270139651622303762018-03-05T10:43:07.963-08:002018-03-05T10:43:07.963-08:00Thanks for the kind words, Anonymous.Thanks for the kind words, Anonymous.Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-26944608728882466112018-03-05T08:31:52.919-08:002018-03-05T08:31:52.919-08:00Very shortly this website will be famous amid all ...Very shortly this website will be famous amid all <br />blogging visitors, due to it's nice articles or reviewsAnonymousnoreply@blogger.com