BERTopic icon indicating copy to clipboard operation
BERTopic copied to clipboard

No keywords and weird topic

Open keyuchen21 opened this issue 3 years ago • 3 comments

  1. There are no keywords for topic 1_cbd_isolate_anxiety_thc
  2. topic 2 seems to have a very weird topic while the keywords looks fine
image

keyuchen21 avatar Jun 02 '22 18:06 keyuchen21

From your dataframe, it seems that you have assigned keywords to the wrong topic by indexing on the previous index. Take the image below:

image

Here, you can see that the words in topic 3, addiction, quit, cravings and addicted actually appear in the keywords of topic 2. This is indicated by the blue rectangles. This is the same for every single topic. The topic words of topic 4 are found with the keywords of topic 3 (red) and the topic words of topic 5 are found with the keywords topic 4 (green).

MaartenGr avatar Jun 02 '22 19:06 MaartenGr

thank you!

But why top 2 is "2____" and also has no keywords with "[,,,,,,,,,]"

keyuchen21 avatar Jun 02 '22 19:06 keyuchen21

Without looking at the documents, my guess would be that the documents in topic 2 are mostly empty or very short such that the resulting topic representations are empty.

MaartenGr avatar Jun 04 '22 05:06 MaartenGr