A Survey of Multi-Label Topic Models

Burkhardt, Sophie; Krämer, Stefan

doi:10.1145/3373464.3373474

Cited by 16 publications

(7 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Alghamdi and Alfalqi (2015) compare 4 topic models (LDA, LSI, PLSA and CTM): this survey studied both their capability in modelling static topics, as well as in detecting topic change over time, highlighting the strengths and weaknesses of each. Burkhardt and Kramer (2019) provide a survey for the adjacent task of multi-label topic models, underlining its challenges and promising directions. Qiang et al (2020) give an extensive performance evaluation of multiple topic models in the context of the Short Text Topic modelling sub-task (e.g.…”

Section: Topic Models Comparisonmentioning

confidence: 99%

Apples to Apples: A Systematic Evaluation of Topic Models

Harrando¹,

Lisena²,

Troncy³

2021

Proceedings of the Conference Recent Advances in Natural Language Processing - Deep Learning for Natural Language Processing Me

View full text Add to dashboard Cite

Section: Topic Models Comparisonmentioning

confidence: 99%

Apples to Apples: A Systematic Evaluation of Topic Models

Harrando¹,

Lisena²,

Troncy³

2021

Proceedings of the Conference Recent Advances in Natural Language Processing - Deep Learning for Natural Language Processing Me

View full text Add to dashboard Cite

“…L-LDA has been widely applied for efficiency and concision, but it constrains the topic distributions in the observed labels that lead to over-focus on them. To alleviate this problem, Dependency-LDA (Rubin et al, 2011) incorporates another topic model to model the observed label correlations, which is deemed to be crucial for multi-label classifiers (Burkhardt and Kramer, 2019b). Another recent improved approach of L-LDA is Twin labeled LDA (Wang et al, 2020b), which employs two sets of parallel topic modeling processes, one incorporates the prior label information by hierarchical Dirichlet distributions, the other models the grouping tags that have prior knowledge about the label correlation.…”

Section: Related Workmentioning

confidence: 99%

“…Statistical topic modeling approaches (Blei, 2012), e.g., Latent Dirichlet Allocation (LDA) (Blei et al, 2003), have been widely applied in the field of data mining, latent data discovery, and document classification (Jelodar et al, 2018). Standard LDA is a completely unsupervised algorithm, and then how to incorporate prior knowledge into the topic modeling procedure is a popular research direction (Burkhardt and Kramer, 2019b;Chen et al, 2019). A major challenge of these LDA customizations is the computational cost of computing the posterior distribution.…”

Section: Introductionmentioning

confidence: 99%

Neural labeled LDA: a topic model for semi-supervised document classification

et al. 2021

View full text Add to dashboard Cite

Recently, some statistical topic modeling approaches based on LDA have been applied in the field of supervised document classification, where the model generation procedure incorporates prior knowledge to improve the classification performance. However, these customizations of topic modeling are limited by the cumbersome derivation of a specific inference algorithm for each modification. In this paper, we propose a new supervised topic modeling approach for document classification problems, Neural Labeled LDA (NL-LDA), which builds on the VAE framework, and designs a special generative network to incorporate prior information. The proposed model can support semi-supervised learning based on the manifold assumption and low-density assumption. Meanwhile, NL-LDA has a consistent and concise inference method while semi-supervised learning and predicting. Quantitative experimental results demonstrate our model has outstanding performance on supervised document classification relative to the compared approaches, including traditional statistical and neural topic models. Specially, the proposed model can support both single-label and multi-label document classification. The proposed NL-LDA performs significantly well on semi-supervised classification, especially under a small amount of labeled data. Further comparisons with related works also indicate our model is competitive with state-of-the-art topic modeling approaches on semi-supervised classification.

show abstract

“…There are different types of topic classification depending on the dataset nature, such as Binary-class and Multi-class, while this paper addresses the scientific articles' topic classification which falls in the type of Multi-labeled Topic classification [2]. In which each article can be assigned to one or more classes.…”

Section: Introductionmentioning

confidence: 99%

SVM transformations for Multi-labeled Topics

El-Sayed¹

2022

Preprint

View full text Add to dashboard Cite

<p>The rapid growth of research papers poses a significant challenge for manual curation and interpretation for categorizing the articles into one or more related topics. The Support Vector Machine (SVM) model addressed this multi-labeled topic classification problem over different datasets, however, transforming the multi-labeled dataset into single-labeled ones to fit the SVM-required modeling plays a vital role in the performance. Applying different transformations over the problem leads to different behavior and accuracy for SVM over datasets.</p> <p>This paper employs the SVM model on a research multi-labeled articles dataset. It addresses different kinds of transformations to measure their behavior. It proposes the Least Class Classifier (LCC) technique that challenges the problem of the imbalanced datasets to achieve an equal chance for the minor classes. The results showed the label powerset transformation achieved the best average accuracy score across all topics classification. Both the label powerset and the binary relevance reached $\approx 90\%$ as the Hamming loss measurement for the fraction of topics that are incorrectly assigned. However, the binary relevance illustrated the best recall and precision balancing as per class classification measurements. Moreover, the proposed LCC technique showed promising results for increasing the recall calculations for the minor class in the imbalanced dataset.</p>

show abstract

A Survey of Multi-Label Topic Models

Cited by 16 publications

References 42 publications

Apples to Apples: A Systematic Evaluation of Topic Models

Apples to Apples: A Systematic Evaluation of Topic Models

Neural labeled LDA: a topic model for semi-supervised document classification

SVM transformations for Multi-labeled Topics

Contact Info

Product

Resources

About