tag:blogger.com,1999:blog-4446292666398344382.post8505500436904308787..comments2024-05-14T20:48:32.117-07:00Comments on Machined Learnings: Even Better Hashtag Similarity VisualizationPaul Mineirohttp://www.blogger.com/profile/05439062526157173163noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-4446292666398344382.post-33365069809064161222011-09-15T13:47:49.366-07:002011-09-15T13:47:49.366-07:00@doug: yes I did. I used Mathematica which lets y...@doug: yes I did. I used Mathematica which lets you rotate and zoom into 3D plots in the notebook interface (not exactly simple: sorry). It seemed to be using the 3rd dimension to improve the cluster differences from each other; but this didn't result in a qualitatively different view. YMMV.<br /><br />I also tried a 3D version using color as the third dimension. This turned out not to be very readable so I didn't post it.Paul Mineirohttps://www.blogger.com/profile/05439062526157173163noreply@blogger.comtag:blogger.com,1999:blog-4446292666398344382.post-48079553023638086032011-09-15T10:04:47.994-07:002011-09-15T10:04:47.994-07:00Did you try generating a 3-D image instead? And t...Did you try generating a 3-D image instead? And the follow-up question: is there a simple rotate/view tool for 3-D matrices out there somewhere -- it really seems to be what t-SNE needs.Doughttps://www.blogger.com/profile/09676401189779405323noreply@blogger.comtag:blogger.com,1999:blog-4446292666398344382.post-42033800109614364792011-06-13T13:25:05.799-07:002011-06-13T13:25:05.799-07:00Yes. The C++ t-SNE implementation does an admirab...Yes. The C++ t-SNE implementation does an admirable job once the data is loaded, actually, but it wants to read the entire dataset as dense vectors into memory before proceeding, so as I tried to scale up it became a problem. The MDS middleman helps with that.<br /><br />I could also have tried the Matlab implementation and fed the distances in, which would have been only 1000x1000 space. However at that point I had a nice workflow going with the c++ version so I ended up doing what I did.Paul Mineirohttps://www.blogger.com/profile/05439062526157173163noreply@blogger.comtag:blogger.com,1999:blog-4446292666398344382.post-3564118728841861392011-06-13T11:53:02.268-07:002011-06-13T11:53:02.268-07:00Thanks for posting this, it's interesting stuf...Thanks for posting this, it's interesting stuff. Did you try just feeding the term frequency vectors directly into t-SNE?Jacob Eisensteinhttps://www.blogger.com/profile/08318716561537991911noreply@blogger.com