ACADSTAFF UGM
| Title | : | Medical subject heading indexing using unsupervised learning |
| Author | : |
Dr. Lukman Heryawan, S.T., M.T. (1) ARDACANDRA S (2) |
| Date | : | 5 2021 |
| Keyword | : | Medical subject heading,Latent Dirichlet allocation, Unsupervised learning,Automatic MeSH indexing,Machine learning Medical subject heading,Latent Dirichlet allocation, Unsupervised learning,Automatic MeSH indexing,Machine learning |
| Abstract | : | Medical subject heading indexing is a process to assign Medical subject heading (MeSH) to the biomedical literatures, such as abstracts or articles. The indexing process was done manually by human MeSH indexer. There is also indexing that was done by machine learning algorithms like supervised learning. Both of processes requires MeSH labels to be assigned manually to the literatures, which is a tedious task and requires a lot of hand-crafted labels from domain experts. In order to do MeSH indexing without manually assigned labels, the unsupervised learning algorithm like LDA (Latent Dirichlet allocation) was performed and analyzed. LDA is kind of unsupervised learning, which can generate topics in a collection of texts, and then automatically classify any individual texts within the collection into a relevant generated topics. In term of MeSH indexing, topic in LDA is a MeSH label. LDA is able to generate MeSH labels from biomedical literatures with good accuracy. Unsupervised learning algorithm like LDA is a promising algorithm to deal with lack of MeSH labels from domain experts. |
| Group of Knowledge | : | |
| Level | : | Internasional |
| Status | : |
Published
|
| No | Title | Action |
|---|