| Full text | |
| Author(s): |
Ferraria, Matheus A.
;
Balbi, Pedro P.
;
de Castro, Leandro N.
Total Authors: 3
|
| Document type: | Journal article |
| Source: | DISTRIBUTED COMPUTING AND ARTIFICIAL INTELLIGENCE, 21ST INTERNATIONAL CONFERENCE; v. 1259, p. 11-pg., 2025-01-01. |
| Abstract | |
This research investigates the challenges and effectiveness of various text representation methods (standard vector, grammar-based, and distributed), when applied to clustering short texts. The study explores Bag-of-Words for standard vector, Linguistic Inquiry and Word Count (LIWC), Part-of-Speech Tagging (POS-Tagging), and the Medical Research Council Psycholinguistic Database (MRC) for grammar-based, and Word2Vec, fastText, Doc2Vec, and SentenceBERT for distributed representations. Utilizing the aiNet bio-inspired clustering algorithm, the results reveal surprising findings, with grammar-based representations demonstrating competitive performance despite their simplicity, while standard vectors exhibit known challenges like high dimensionality. The study contributes insights into the properties of different text representations, providing a foundation for optimizing their application in clustering tasks with short and informal texts. (AU) | |
| FAPESP's process: | 21/11905-0 - Center of Science, Technology and Development for innovation in Medicine and Health: inLab.iNova |
| Grantee: | Giovanni Guido Cerri |
| Support Opportunities: | Research Grants - Science Centers for Development |