How does Latent Dirichlet Allocation (LDA) works in unsupervised mood to extract topic from data set ?
LDA is unsupervised by nature, hence it does not need predefined dictionaries.
This means it finds topics automatically, but you cannot control the kind of topics it finds.
That’s right that LDA is an unsupervised method. However, it could be extended to a supervised one.
For more details about the LDA method and its extensions, you can consult these links:
https://en.wikipedia.org/wiki/Latent_Dirichlet_allocation
https://www.cs.princeton.edu/~blei/papers/BleiMcAuliffe2007.pdf
http://jmlr.org/proceedings/papers/v28/zhai13.pdf
http://papers.nips.cc/paper/4094-word-features-for-latent-dirichlet-allocation.pdf
Hi,
try to look in these papers, look interesting for topic modeling. Also the code and datasets are available on author's web site.
Zhiyuan Chen and Bing Liu. Mining Topics in Documents: Standing on the Shoulders of Big Data.. Proceedings of the 20th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2014), August 24-27, New York, USA.
Zhiyuan Chen and Bing Liu. Topic Modeling using Topics from Many Domains, Lifelong Learning and Big Data. Proceedings of the 31st International Conference on Machine Learning (ICML 2014), June 21-26, Beijing, China
- Badges
- Science method