A Novel Method of Fuzzy Topic Modeling based on Transformer Processing
This work addresses the challenge of improving topic modeling interpretability for market trend monitoring, though it appears incremental by combining existing fuzzy clustering and transformer techniques.
The authors tackled the problem of traditional topic modeling methods like LDA producing non-intuitive results and requiring manual topic number selection by proposing a fuzzy topic modeling method based on soft clustering and transformer embeddings. In a press release monitoring application, their method yielded more natural results compared to LDA.
Topic modeling is admittedly a convenient way to monitor markets trend. Conventionally, Latent Dirichlet Allocation, LDA, is considered a must-do model to gain this type of information. By given the merit of deducing keyword with token conditional probability in LDA, we can know the most possible or essential topic. However, the results are not intuitive because the given topics cannot wholly fit human knowledge. LDA offers the first possible relevant keywords, which also brings out another problem of whether the connection is reliable based on the statistic possibility. It is also hard to decide the topic number manually in advance. As the booming trend of using fuzzy membership to cluster and using transformers to embed words, this work presents the fuzzy topic modeling based on soft clustering and document embedding from state-of-the-art transformer-based model. In our practical application in a press release monitoring, the fuzzy topic modeling gives a more natural result than the traditional output from LDA.