User Tools

Site Tools


latentartists

====== Differences ====== This shows you the differences between two versions of the page.

Link to this comparison view

latentartists [2013/06/30 16:53]
ben
latentartists [2013/06/30 17:09] (current)
ben
Line 7: Line 7:
  
  
-Assume a fixed vocabulary $V$, which in our experiments is a company internal ​list+Assume a fixed vocabulary $V$, which in our experiments is a list compiled by the Echonest
 of music related multiword terms. of music related multiword terms.
  
Line 69: Line 69:
 during training. during training.
  
-{{::1000.ps|}}+== Hottt 1000 ==
  
-{{::full.ps|}}+{{::​1000.jpg?​600|}} 
 + 
 +== Full Dataset == 
 + 
 +{{::full.jpg?600|}}
  
 The results do not support our hypothesis that taking uncertainty into account would create a more robust notion of similarity. ​ The results do not support our hypothesis that taking uncertainty into account would create a more robust notion of similarity. ​
-While both methods ​effectively ​capture the artist similars ​in the Echonest ​database, the area under the ROC curve is clearly+While both methods ​clearly ​capture the information ​in the Echonest ​artist similar lists, the area under the ROC curve is clearly
 greater for the simple Euclidean distance based approach.  ​ greater for the simple Euclidean distance based approach.  ​
  
 The reason that the experimental results do not match our intuition is unclear. ​ One possibility is that KL divergence The reason that the experimental results do not match our intuition is unclear. ​ One possibility is that KL divergence
 is not an appropriate metric for similarity.  ​ is not an appropriate metric for similarity.  ​
- 
-{{::​nice-bullseye-o.gif?​300|}} 
- 
- 
  
  
latentartists.1372625638.txt.gz · Last modified: 2013/06/30 16:53 by ben