I have a lot of outliers in my project, I want to get the probability of each sentence belonging to each topic, If I use topics,probs= topic_model.fit_transform(sentences,embeddings=embeddings) the probs obtained by this method will contain every sentence (including outliers). Is there any way to extract these outliers, and do these outliers also have topic probability proportions

[superseanyoung]

topic_model = BERTopic(embedding_model=transformer_model,
min_topic_size=3,
verbose=True,
umap_model=umap_model,
hdbscan_model=hdbscan_model,
ctfidf_model=ctfidf_model,
representation_model=representation_model,
#top_n_words=10,
#min_topic_size=10,
#nr_topics=None,
#low_memory=False,
calculate_probabilities=True
)
topics,probs= topic_model.fit_transform(sentences,embeddings=embeddings)

[superseanyoung]

In the obtained probs, I do not want to get sentences belonging to outliers, how should I post them

[MaartenGr]

Have you checked out the documentation of BERTopic? You should be able to find the solution there.

To help you out, please refer to the FAQ about this specific question.

[superseanyoung]

I update the topic after reducing the outliers, whether this will change the previous probs

[MaartenGr]

The probabilities should stay the same after reducing outliers if you do not merge topics together.