Michael Röder

Hi Velcin, thanks for using Palmetto and for sharing these interesting examples. The bad topic seems to contain words that are very rare. Maybe this is a special case in...

@provalis thanks for sharing your results. Can you please describe your preprocessing process? @Velcin from my point of view, the problem is caused by the difference in the preprocessing of...

Did you used your own index for the numbers reported above? I thought you have used the webservice.

Finally, I had some time to take a deeper look into the two examples of Velcin. The probabilities clearly support your impression, that the bad topic should have a very...

Hi Velcin, I think that the major problem of the NPMI implementation of Palmetto is the epsilon (e). As described in the paper, the NPMI is implemented as ``` NPMI(W',W*)=...

Yes, after making sure that the removal of the epsilon does not create other side effects I would like to implement NPMI* and C_V* (maybe with different names) and recompute...

You did everything correct :+1: #### 1. Dockerfile There should be no problem after updating to a newer tomcat 8 version. #### 2. Including own index in docker container I...

In general, this behavior is expected if you try to use many threads that evaluate different topics in parallel. The problem is that window-based coherence measures need to know the...

Your setup looks good and should work. I am just wondering why you have `-Xmx3072m` in the options as it limits the server to use not more than 3GB of...

That depends a lot on how you actually use it. The main bottleneck is reading the index. Using a GPU instead of a CPU does most probably not help since...