Jämför metoder

	Svagt övervakad LDA-ämnesskapare	Svag övervakad BERT-baserad klassificering
Ämnesområde	Djupinlärning	Djupinlärning
Familj	Machine learning	Machine learning
Ursprungsår	2009–2012	2017–2020
Upphovsperson	Jagarlamudi et al.; Andrzejewski et al.	Multiple (Ratner et al. for weak supervision framework; Meng et al. for BERT integration)
Typ	Probabilistic generative model with weak supervision	Weakly supervised fine-tuning of pre-trained language model
Ursprungskälla	Jagarlamudi, J., Daume III, H., & Udupa, R. (2012). Incorporating Lexical Priors into Topic Models. Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2012), pp. 204–213.	Meng, Y., Zhang, Y., Huang, J., Xiong, C., Ji, H., Zhang, C., & Han, J. (2020). Text Classification Using Label Names Only: A Language Model Self-Training Approach. Proceedings of EMNLP 2020, 9006–9017.
Alias	WS-LDA, Guided LDA, Seeded LDA, Constrained LDA	WS-BERT, BERT with weak supervision, label-efficient BERT classification, noisy-label BERT fine-tuning
Närliggande	6	6
Sammanfattning	Weakly Supervised LDA is an extension of Latent Dirichlet Allocation that incorporates lightweight human guidance — typically keyword seeds or must-link/cannot-link constraints — into the Dirichlet priors, steering learned topics toward domain-meaningful themes without requiring fully labeled documents. It sits between fully unsupervised LDA and supervised classification, making it well-suited to situations where labeling thousands of documents is impractical.	Weakly supervised BERT-based classification adapts BERT to text classification tasks when only noisy, heuristic, or programmatically generated labels are available instead of clean human annotations. It combines weak supervision frameworks — such as labeling functions and data programming — with BERT's pre-trained language representations to achieve robust classification without expensive hand-labeling.