Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

Fine-Tuning Topics through Weighting Aspect Keywords

T0 review · 4 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Expert-informed aspect weighting sharpens LDA topic models by overlaying curated keyword sets onto an unsupervised baseline.

desk verdict A transparent, reproducible guided-topic-modeling pipeline whose evaluation does not yet separate expert input from mechanical keyword overlap; fixable, but not acceptable as is. read the letter →

arxiv 2502.08496 v2 pith:366NRGRB submitted 2025-02-12 cs.IR cs.LG

Ali Nazari , Michael Weiss This is my paper · ORCID
classification cs.IRcs.LG
keywords TopicmodelingExpert-informedkeywordsWeightedaspectAdaptivelearningDocumentrelevanceSupervisedclusteringQuantumcommunicationLatentDirichletAllocation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Standard topic models like Latent Dirichlet Allocation (LDA) are unsupervised and static: they reflect dominant word frequencies, not what domain experts consider important. This paper proposes an 'aspect-weighted' overlay: the practitioner defines a few aspects (e.g., Cryptography, QKD), represents each by its top 50 TF-IDF-weighted keywords from expert sources such as conference proceedings, scores every document by cosine similarity to these aspect vectors, and reassigns documents to the best-matching aspect-topic. Iterating this reassignment until few documents move is claimed to increase topic coherence, raise the visibility of rare strategic terms, and align topics with expert priorities. Applied to a quantum-communication corpus, the method reclassified documents into more thematically coherent clusters and tracked a shift in QCrypt conference papers from foundational theory toward implementation and classical–quantum interoperability. The paper's core claim is that expert-informed aspect weighting is a simple yet effective way to boost the relevance and clarity of topic modeling outputs, and that applying it iteratively helps align the model with shifting domain signals.

What carries the argument

The machinery is the 'aspect-topic model': an LDA topic model overlaid with expert-defined aspect vectors. Each aspect $A_i$ is a set of the top 50 TF-IDF-weighted keywords from an expert corpus (e.g., conference agendas), and each topic $T_j$ is its top 100 LDA terms with probabilities. Relevance between aspect $i$ and topic $j$ is computed by the weighted stem-overlap score in Eq. (2), which is nonzero only when an aspect keyword and a topic term share the same stem; documents are then aligned to aspect-topics by cosine similarity (Eq. 4), and the aspect centroids are updated iteratively until document reassignment stabilizes. This overlay is what lets expert knowledge steer topic boundaries without re-training the base LDA model.

What would settle it

Take an aspect concept that an expert describes only with words absent from the topic's top-100 stems (e.g., 'authenticated' instead of 'secur'). Under Eq. (2) the relevance score is exactly zero; if a human rater judges the document to be thematically relevant to the aspect, the central claim that the scoring reflects expert alignment would be refuted. A concrete check: replace one aspect's keyword list with synonyms sharing no stems and observe whether document-to-topic reassignments stay stable.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that an LDA topic model can be fine-tuned without retraining by injecting expert-defined aspect vectors. Each aspect is a weighted keyword list, built by running TF-IDF on text from authoritative sources (here, Quantum.Tech conference material), and each topic's word distribution is compared with those aspect vectors using a weighted stem-overlap score $R_{ij} = \sum_{k,l} w_{ik} v_{jl}\, \mathrm{sim}(a_{ik}, t_{jl})$, with similarity 1 only when stems match. Documents are then re-assigned to the aspect-topic with the highest cosine similarity, and the process loops until document reassignment nearly stops. In the quantum cryptography case study, the paper reports that the aspect-weighted model increased intra-cluster similarity for topic T19 from 0.61 to 0.74, cut the reassignment rate from 28% to 4.2%, and produced relevance scores up to 0.506 between the Cryptography aspect and T19. Testing on QCrypt 2023 and 2024 papers, it finds the model adapts to a discourse shift from foundational theory to implementation. The conclusion is that expert-informed aspect weighting is a simple yet effective way to boost the relevance and clarity of topic modeling outputs, and that iterative application helps align the model with shifting domain signals.

Load-bearing premise

The method assumes that exact stem overlap between the top-50 TF-IDF aspect keywords and the top-100 LDA topic terms is a good enough measure of semantic alignment: if real synonyms or paraphrases are missed, the reassignment will not actually reflect expert aspects.

Editorial extensions

If this is right

  • A single overlay can update topic models with new expert knowledge: adding an aspect from a later conference re-weights topic–document relationships without retraining the base LDA model.
  • Rare but strategically important terms become more visible: weighting low-frequency expert keywords raises their influence on document clustering, which is useful for tracking post-quantum cryptography and similar niches.
  • The framework yields cleaner, more interpretable clusters: in the case study, documents with weak cryptographic relevance were re-assigned away from cryptography topics, shifting weights such as Doc9's T19 weight from 0.255 to 0.676.
  • Topic evolution can be tracked: the two-iteration comparison on QCrypt 2023/2024 papers shows the model moving from foundational protocols to classical–quantum interoperability.
  • The method stays transparent and auditable: each step—keyword lists, weights, relevance scores, and heatmaps—can be inspected, unlike black-box summarizers.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The success of the framework likely depends on the semantic coverage of the expert keyword lists; if the expert source is too small, top-50 TF-IDF terms may miss important concepts—an extension would be to mine keywords from multiple sources or use synonym expansion.
  • The proposed convergence criterion (few documents moving between clusters) could be made a formal fixed-point condition, which would make the stopping rule more principled than a heuristic threshold.
  • Beyond quantum communication, the same overlay should be testable in other specialist domains with an authoritative conference corpus (e.g., AI safety, synthetic biology), where weak signals and expert priority shifts matter.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes an Aspect-Weighted Topic Modeling framework that integrates expert-curated aspect keywords into an LDA-based pipeline. The framework proceeds through four phases: baseline topic modeling with LDA and hierarchical clustering, construction of weighted aspect vectors from expert conference sources via TF-IDF, document-to-aspect alignment using cosine similarity with exact stem matching, and iterative refinement until document reassignment stabilizes. The method is applied to a quantum communication corpus of 1,048 documents, with QCrypt 2023 and 2024 papers used as temporal test sets. The authors report improved intra-cluster similarity, reduced reassignment rates, better alignment with expert themes, and the ability to detect a shift from theoretical to implementation-focused cryptography topics. The paper positions the contribution as a transparent, modular, and adaptable alternative to static unsupervised topic models.

Significance. If the claimed benefits were convincingly established, the framework would offer a practical way to inject expert knowledge into topic modeling without full retraining, with potential value for technology monitoring and weak-signal detection in fast-moving domains. The manuscript has several strengths: the method is described in detail with formal notation, pseudocode, and a public repository; the design-science framing provides traceability; and the authors clearly acknowledge limitations in Section 7.5. However, the significance is currently undercut by the evaluation design: the reported improvements lack a non-expert baseline, the temporal adaptability evidence is circular because the aspect definitions and test papers share the same years, and the claimed coherence gains are not measured with any coherence metric. These issues directly affect the central claim that expert weighting, rather than the reassignment rule itself, causes the observed improvements.

major comments (4)
  1. [Section 4.6, Eq. (2); Section 6] The similarity measure sim(a_ik, t_jl) in Eq. (2) is 1 only when the aspect stem and topic stem are identical, and Eq. (4) then uses cosine similarity over the shared vocabulary. With this construction, any keyword set that overlaps document vocabulary will mechanically increase within-cluster similarity for documents containing those stems. The evaluation in Section 6 reports intra-cluster similarity rising from 0.61 to 0.74 and reassignment falling from 28% to 4.2%, but it provides no baseline with random keywords, non-expert keywords, or a simple TF-IDF-only keyword set. Therefore the central claim that expert-informed weighting, rather than the exact-stem reassignment rule itself, drives the improvement is not established.
  2. [Section 4.9; Section 5; Section 7.1] The temporal adaptability evaluation is circular. Section 4.9 states that Aspect 1 was defined from the 2023 Quantum.Tech report and Aspect 2 from the 2024 report, while Section 5 tests the framework on QCrypt 2023 and QCrypt 2024 papers. The model aligns 2023 papers with Aspect 1 and 2024 papers with Aspect 2, which is essentially fitting the output to the input year labels. The observed 'shift from theoretical foundations to implementation challenges' therefore reflects the changed keyword lists, not model adaptation to evolving domain discourse. To support the adaptability claim, the authors need a held-out temporal split, such as defining aspects on years strictly before the test years, or a control in which aspect keywords are shuffled across years.
  3. [Section 6; Section 4.4] Although the paper repeatedly claims improved topic coherence and interpretability, no coherence metric is reported for the refined models. The C-V scores mentioned in Section 4.4 are used only to select the number of primary topics in the baseline LDA; they are not computed for the aspect-weighted topics. The evaluation evidence consists of keyword tables, heatmaps, and the intra-cluster similarity/reassignment numbers, but these are reported without error bars, confidence intervals, or statistical tests. Without a formal coherence measure and a variance estimate, the coherence and stability claims are not quantitatively supported.
  4. [Section 7.5; Section 8] The authors acknowledge in Section 7.5 that cosine similarity 'might miss deeper semantic relationships' and that topic selection was manual, yet Section 8 concludes that aspect weighting 'is a simple yet effective way to boost the relevance and clarity of topic modeling outputs.' The limitation statement does not address the more fundamental issue that the evaluation cannot distinguish expert-informed weighting from the mechanical effect of stem-overlap reassignment. A revision should add a non-expert control condition and a proper semantic similarity measure, or the conclusions should be substantially softened to describe only the exact-stem re-ranking behavior.
minor comments (4)
  1. [Section 4.4, Figures 5 and 6] Figure 5 is referenced both as 'Coherence Scores for Various Numbers of Primary Topics' and later in Section 4.4 as 'Silhouette Scores for Various Numbers of SubTopics'; the figure numbering and captions appear inconsistent and should be corrected.
  2. [Section 4.2, Figure 3] The caption of Figure 3 and the component list in the text use slightly different names for the same components (e.g., 'Document Realignment via Semantic Scoring' versus 'Supervised Clustering and Refinement'), which may confuse readers.
  3. [Section 9 (Availability of data and materials)] The availability statement gives the repository https://github.com/alinazari1/FineTuning/blob/main/, while Section 4.6 references a different URL (https://github.com/AspectTopicModels). Please verify and unify the repository link, and ensure the data referenced as an accompanying .xlsx file is actually accessible.
  4. [Throughout] There are numerous typos and grammatical issues, such as 'these stands together emphasize' in Section 2, 'it is making it well-suited' in Section 7.3, and 'the authors weighted the documents more with cryptography keywords' in Section 4.6. A thorough language edit is needed.

Circularity Check

3 steps flagged · score 8.0 of 10

Aspect definitions, exact-stem matching, and the QCrypt validation sources make the reported alignment gains and temporal shift largely self-fulfilling.

  1. self definitional [Section 4.9 (Formal Notation) and Section 5 (Results)]
    "Specifically, the 2022 Quantum.Tech report was used to extract nine core aspects, while the 2023 and 2024 cryptography reports were later used to define Aspect 1 and Aspect 2, respectively, for the two iterations of topic model refinement. ... Aspect 1 ... aligned more closely with theoretical papers from QCrypt 2023 papers ... Aspect 2 ... showed stronger relevance to QCrypt 2024 papers ... The similarity measure equals 1 only when both term stems are the same."

    Aspect 1 and Aspect 2 are built from 2023 and 2024 conference texts, respectively; the test documents are QCrypt 2023 and QCrypt 2024 papers; and relevance is computed by exact-stem overlap (sim=1 only for identical stems). Any year-specific vocabulary in the test papers will therefore reproduce the year labels: 2023 papers match the 2023-derived aspect and 2024 papers match the 2024-derived aspect. The claimed shift from theoretical foundations to implementation challenges is an artifact of feeding year-dependent keyword lists into the matching rule, not an independent detection of domain evolution.

  2. fitted input called prediction [Section 4.7 (Iterative Refinement) and Section 6 (Evaluation)]
    "After each round, aspect centroids are updated and documents reassessed using cosine similarity. Documents aligning more strongly with a new aspect-topic centroid are reassigned. This process continued until convergence... stabilization of within-cluster similarity scores. ... the topic stability improved between iterations such as in T19, as shown by the reduced reassignment rate (28% to 4.2%) and increased intra-cluster similarity (0.61 to 0.74)."

    Documents are reassigned to the centroid that maximizes cosine similarity, and convergence is defined by stabilization of within-cluster similarity. Therefore the reported increase in intra-cluster similarity (0.61 to 0.74) and the low final reassignment rate are direct consequences of the optimization objective, not independent evidence that expert weighting improved coherence. The outcome metric is the same function being optimized during refinement.

1 more flagged steps
  1. self definitional [Section 6.2 (Validation Approach)]
    "Selected external sources, QCrypt 2023 and 2024 conference proceedings, were used as expert-informed proxies for validate emerging technologies from the models. These sources helped define the aspects and guided the alignment logic, ensuring that the method reflected real-world subthemes relevant to early signal detection and domain-specific mapping."

    The validation corpus is used to define the aspects and alignment logic, and then the same corpus, QCrypt 2023 and 2024, is used to validate the framework. The supposedly external, time-stamped test corpora are not external to the construction: the years and vocabulary used to shape the model are the same years and vocabulary used to demonstrate success, so the validation is circular.

full rationale

The central demonstration is not independent of its inputs. Aspect 1 and Aspect 2 are year-labeled keyword vectors built from 2023 and 2024 sources; QCrypt 2023 and 2024 test papers are matched by exact-stem cosine overlap, so the detected temporal shift is essentially the year label propagating through Eq. 2. The improvement in intra-cluster similarity is the same cosine objective used to reassign documents, and Section 6.2 explicitly says the QCrypt test sources helped define the aspects. No external benchmark, non-expert baseline, or randomized control is provided. The framework has some independent machinery, such as the LDA baseline, 39 subtopics, and iterative updates, but the paper's headline claims of expert-driven alignment and shift detection are constructed by the evaluation design rather than demonstrated. There is no load-bearing self-citation chain.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The framework's central mechanism depends on several unvalidated choices: the number of keywords, the unspecified boosting weights, the manual topic selection, and the exact-match similarity measure. The paper acknowledges the manual selection and the crude similarity measure in its Limitations.

free parameters (5)
  • Number of aspect keywords per aspect = 50
    Section 4.5.2 states 'top 50 TF-IDF-weighted keywords per aspect' are used. This arbitrary choice determines the aspect vectors.
  • Number of topic keywords per topic = 100
    Section 4.6 defines topics with m=100 keywords, which affects the relevance score in Eq. (2).
  • Elevated weights for expert-flagged low-frequency terms = unstated
    Section 4.5.2 says expert-flagged rare terms are assigned 'elevated weights' but no formula or value is given. This ad hoc adjustment influences results.
  • Convergence threshold = unspecified
    Section 4.7 defines convergence as 'fewer documents shifting' and 'stabilization' without numeric criteria, making the stopping point subjective.
  • Manual selection of five cryptography topics = T19, T21, T32, T33, T39 (varies)
    Section 7.5 acknowledges the selection of five main topics for the cryptography aspect was manual, introducing selection bias into the reported results.
assumptions (5)
  • standard math LDA with chosen hyperparameters produces valid latent topics
    Section 4.4 uses LDA with C-V coherence to choose K=6. This assumes standard LDA assumptions hold for the corpus.
  • domain assumption TF-IDF on aspect corpora yields representative keyword vectors
    Section 4.5.2 uses TF-IDF to select top keywords from conference materials. This assumes these keyword lists capture the meaning of each aspect.
  • domain assumption Quantum.Tech conference web content is a valid proxy for expert knowledge
    Section 4.5.1 states aspect corpora are built from Quantum.Tech materials as a stand-in for expert input. If this source is not representative, the alignment is not actually expert-guided.
  • ad hoc to paper Exact stem matching is an adequate similarity measure between aspect and topic terms
    Section 4.6, Eq. (2) sets sim(a_ik, t_jl)=1 only for identical stems. This strong assumption underlies the relevance scores and document reassignment.
  • domain assumption QCrypt 2023 and 2024 papers are suitable external test sets for evaluating temporal shifts
    Section 5 uses these conference papers to validate the framework. The choice of test set is not justified beyond being from the same field.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fine-Tuning Topics through Weighting Aspect Keywords." pith.science (2026). https://pith.science/paper/366NRGRB

@misc{pith2026250208496,
  author       = {Pith},
  title        = {Pith review of: Fine-Tuning Topics through Weighting Aspect Keywords},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/366NRGRB}},
  note         = {Machine review of arXiv:2502.08496}
}
read the original abstract

Organizations face growing challenges in deriving meaningful insights from vast amounts of specialized text data. Conventional topic modeling techniques are typically static and unsupervised, making them ill-suited for fast-evolving fields like quantum cryptography. These models lack contextual awareness and cannot easily incorporate emerging expert knowledge or subtle shifts in subdomains. Moreover, they often overlook rare but meaningful terms, limiting their ability to surface early signals or align with expert-driven insights essential for strategic understanding. To tackle these gaps, we employ design science research methodology to create a framework that enhances topic modeling by weighting aspects based on expert-informed input. It combines expert-curated keywords with topic distributions iteratively to improve topic relevance and document alignment accuracy in specialized research areas. The framework comprises four phases, including (1) initial topic modeling, (2) expert aspect definition, (3) supervised document alignment using cosine similarity, and (4) iterative refinement until convergence. Applied to quantum communication research, this method improved the visibility of critical but low-frequency terms. It also enhanced topic coherence and aligned topics with the cryptographic priorities identified by experts. Compared to the baseline model, this framework increased intra-cluster similarity. It reclassified a substantial portion of documents into more thematically accurate clusters. Evaluating QCrypt 2023 and 2024 conference papers showed that the model adapts well to changing discussions, marking a shift from theoretical foundations to implementation challenges. This study illustrates that expert-guided, aspect-weighted topic modeling boosts interpretability and adaptability.

Figures

Figures reproduced from arXiv: 2502.08496 by the authors.

Figure 1
Figure 1. Paper S1 S2 S3 S4 S5 Relevance [8] ✓ ✓ Clustering Massive-Categories and Complex Documents via Graph Convolutional Network. [41] ✓ Semi supervised Fuzzy Clustering with Partition Information of Subsets. [22] ✓ ✓ Research proposal content extraction using natural language processing and semi￾supervised clustering: A demonstration and comparative analysis. [21] ✓ ✓ Document clustering with dual supervision through fea… view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Exploring the Technology Landscape through Topic Modeling, Expert Involvement, and Reinforcement Learning

    cs.LG 2025-01 reject novelty 4.0 of 10

    The paper combines LDA topic models, TF-IDF expert keywords, and a four-metric Q-learning reward to select topics, then evaluates selection using the same conference papers that generated the rewards.

Reference graph

Works this paper leans on

65 extracted references · 62 canonical work pages · cited by 1 Pith paper

  1. [1]

    Latent dirichlet allocation

    Blei DM, Ng AY, Jordan MI. Latent dirichlet allocation. Journal of machine Learning research. 2003;3(Jan):993–1022

    work page 2003

  2. [2]

    BERTopic: Neural topic modeling with a class-based TF-IDF procedure

    Grootendorst M. BERTopic: Neural topic modeling with a class-based TF-IDF procedure. arXiv; 2022, Available from: http://arxiv.org/abs/2203.05794

    arXiv 2022

  3. [3]

    The application of text mining methods in innovation research: current state, evolution patterns, and development priorities

    Antons D, Grünwald E, Cichy P, Salge TO. The application of text mining methods in innovation research: current state, evolution patterns, and development priorities. R&D Management. 2020;50(3):329–51

    work page 2020

  4. [4]

    Evolution of topics and trends in emerging research fields: multiple analyses with entity linking, Mann–Kendall test and burst methods in cloud computing

    Coccia M, Roshani S. Evolution of topics and trends in emerging research fields: multiple analyses with entity linking, Mann–Kendall test and burst methods in cloud computing. Scientometrics. 2024 Sep;129(9):5347–71

    work page 2024

  5. [5]

    Short Text Clustering with a Deep Multi-embedded Self-supervised Model

    Zhang K, Lian Z, Li J, Li H, Hu X. Short Text Clustering with a Deep Multi-embedded Self-supervised Model. In: Farkaš I, Masulli P, Otte S, Wermter S, editors. Artificial Neural Networks and Machine Learning – ICANN 2021. Cham: Springer International Publishing; 2021. p. 150–61. (Lecture Notes in Computer Science; vol. 12895). Available from: https://link...

    work page doi:10.1007/978-3-030-86383-8_12 2021

  6. [6]

    Document clustering with dual supervision through feature reweighting

    Hu Y, Milios EE, Blustein J. Document clustering with dual supervision through feature reweighting. Comput Intell. 2016 Aug;32(3):480–513

    work page 2016

  7. [7]

    Hyperspherical Fuzzy clustering for online document categorization

    Mei JP, Wang Y. Hyperspherical Fuzzy clustering for online document categorization. In: 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE). Vancouver, BC, Canada: IEEE; 2016

    work page 2016

  8. [8]

    Clustering massive-categories and complex documents via graph convolutional network

    Zhao Q, Yang J, Wang Z, Chu Y, Shan W, Tuhin IAK. Clustering massive-categories and complex documents via graph convolutional network. In: Knowledge Science, Engineering and Management. Cham: Springer International Publishing; 2021. p. 27–39. (Lecture notes in computer science)

    work page 2021

Show all 65 references
  1. [9]

    The integrative domain of foresight and competitive intelligence and its impact on R&D management: Integrative domain of foresight and competitive intelligence

    Calof J, Smith J. The integrative domain of foresight and competitive intelligence and its impact on R&D management: Integrative domain of foresight and competitive intelligence. R&D Management. 2009 Dec 16;40(1):31–9

    2009

  2. [10]

    Linking technology intelligence to open innovation

    Veugelers M, Bury J, Viaene S. Linking technology intelligence to open innovation. Technological forecasting and social change. 2010;77(2):335–43

    2010

  3. [11]

    Systematic Mapping Studies in Software Engineering

    Petersen K, Feldt R, Mujtaba S, Mattsson M. Systematic Mapping Studies in Software Engineering. In

  4. [12]

    Guidelines for conducting systematic mapping studies in software engineering: An update

    Petersen K, Vakkalanka S, Kuzniarz L. Guidelines for conducting systematic mapping studies in software engineering: An update. Information and Software Technology. 2015 Aug;64:1–18

    2015

  5. [13]

    A correlated topic model of science

    Blei DM, Lafferty JD. A correlated topic model of science. 2007; Available from: https://projecteuclid.org/journals/annals-of-applied-statistics/volume-1/issue-1/----Custom-HTML---- A/10.1214/07-AOAS114.short

    2007 doi

  6. [14]

    An empirical study on innovation ecosystem, technological trajectory transition, and innovation performance

    Sun Y, Li L, Chen Y, Kataev MY. An empirical study on innovation ecosystem, technological trajectory transition, and innovation performance. Journal of Global Information Management (JGIM). 2021;29(4):148–71

    2021

  7. [15]

    Linguistic regularities in continuous space word representations

    Mikolov T, Yih W tau, Zweig G. Linguistic regularities in continuous space word representations. In: Proceedings of the 2013 conference of the north american chapter of the association for computational linguistics: Human language technologies. 2013. p. 746–51. Available from:...

    2013

  8. [16]

    Term-weighting approaches in automatic text retrieval

    Salton G, Buckley C. Term-weighting approaches in automatic text retrieval. Information processing & management. 1988;24(5):513–23

    1988

  9. [17]

    Incremental fuzzy clustering for document categorization

    Mei JP, Wang Y, Chen L, Miao C. Incremental fuzzy clustering for document categorization. In: 2014 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE). Beijing, China: IEEE; 2014. 33

    2014

  10. [18]

    Probabilistic word selection via topic modeling

    Zhuang Y, Gao H, Wu F, Tang S, Zhang Y, Zhang Z. Probabilistic word selection via topic modeling. IEEE Transactions on Knowledge and Data Engineering. 2014;27(6):1643–55

    2014

  11. [19]

    Incorporating lexical priors into topic models

    Jagarlamudi J, Daumé III H, Udupa R. Incorporating lexical priors into topic models. In: Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics. 2012. p. 204–

    2012

  12. [20]

    Seed-guided topic model for document filtering and classification

    Li C, Chen S, Xing J, Sun A, Ma Z. Seed-guided topic model for document filtering and classification. ACM Trans Inf Syst. 2019 Jan;37(1):1–37

    2019

  13. [21]

    Available from: https://aclanthology.org/E12-1021.pdf

  14. [22]

    Research proposal content extraction using natural language processing and semi-supervised clustering: A demonstration and comparative analysis

    Knisely BM, Pavliscsak HH. Research proposal content extraction using natural language processing and semi-supervised clustering: A demonstration and comparative analysis. Scientometrics. 2023 May;128(5):3197–224

    2023

  15. [23]

    Document Clustering With Dual Supervision Through Feature Reweighting

    Hu Y, Milios EE, Blustein J. Document Clustering With Dual Supervision Through Feature Reweighting. Computational Intelligence. 2016 Aug;32(3):480–513

    2016

  16. [24]

    Learning to cluster documents into workspaces using large scale activity logs

    Kong W, Bendersky M, Najork M, Vargo B, Colagrosso M. Learning to cluster documents into workspaces using large scale activity logs. In: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York, NY, USA: ACM; 2020

    2020

  17. [25]

    Automatic constraints generation for semisupervised clustering: experiences with documents classification

    Diaz-Valenzuela I, Loia V, Martin-Bautista MJ, Senatore S, Vila MA. Automatic constraints generation for semisupervised clustering: experiences with documents classification. Soft Comput. 2016 Jun;20(6):2329– 39

    2016

  18. [26]

    Heterogeneity of optimal balance between exploration and exploitation:the moderating roles of firm technological capability and industry alliance network position

    Seo E, Song ,Jaeyong, and Jin C. Heterogeneity of optimal balance between exploration and exploitation:the moderating roles of firm technological capability and industry alliance network position. Industry and Innovation. 2023 Apr 21;30(4):423–51

    2023

  19. [27]

    The Interplay Between Exploration and Exploitation

    Gupta AK, Smith KG, Shalley CE. The Interplay Between Exploration and Exploitation. AMJ. 2006 Aug;49(4):693–706

    2006

  20. [28]

    Intention-guided deep semi-supervised document clustering via metric learning

    Jingnan L, Chuan L, Ruizhang H, Yongbin Q, Yanping C. Intention-guided deep semi-supervised document clustering via metric learning. J King Saud Univ - Comput Inf Sci. 2023 Jan;35(1):416–25

    2023

  21. [29]

    Enhancing neural topic model with multi-level supervisions from seed words

    Lin Y, Gao X, Chu X, Wang Y, Zhao J, Chen C. Enhancing neural topic model with multi-level supervisions from seed words. In: Findings of the Association for Computational Linguistics: ACL 2023. Stroudsburg, PA, USA: Association for Computational Linguistics; 2023

    2023

  22. [30]

    Design Science in Information Systems Research

    Hevner AR, March ST, Park J, Ram S. Design Science in Information Systems Research. MIS Quarterly. 2004;28(1):75–105

    2004

  23. [31]

    Anticipating Future Innovation Pathways Through Large Data Analysis

    Diam TU, Chiavetta D, Porter AL, Saritas O, editors. Anticipating Future Innovation Pathways Through Large Data Analysis. Cham: Springer International Publishing; 2016. (Innovation, Technology, and Knowledge Management). Available from: http://link.springer.com/10.1007/978-3-3...

    2016 doi

  24. [32]

    Secure Quantum Communication Technologies and Systems: From Labs to Markets

    Cavaliere F, Prati E, Poti L, Muhammad I, Catuogno T. Secure Quantum Communication Technologies and Systems: From Labs to Markets. Quantum Reports. 2020 Jan 22;2(1):80–106

    2020

  25. [33]

    A Design Science Research Methodology for Information Systems Research

    Peffers K, Tuunanen T, Rothenberger MA, Chatterjee S. A Design Science Research Methodology for Information Systems Research. Journal of Management Information Systems. 2007 Dec;24(3):45–77

    2007

  26. [34]

    Present landscape of quantum computing

    Hassija V, Chamola V, Saxena V, Chanana V, Parashari P, Mumtaz S, et al. Present landscape of quantum computing. IET Quantum Communication. 2020 Dec;1(2):42–8

    2020

  27. [35]

    Quantum Communications in Future Networks and Services

    Manzalini A. Quantum Communications in Future Networks and Services. Quantum Reports. 2020 Mar 11;2(1):221–32

    2020

  28. [36]

    Introduction to information retrieval

    Manning CD. Introduction to information retrieval. Cambridge university press; 2008. Available from: http://diglib.globalcollege.edu.et:8080/xmlui/bitstream/handle/123456789/1096/Manning_introduction_to_i nformation_retrieval.pdf?sequence=1&isAllowed=y

    2008

  29. [37]

    Active Learning Literature Survey

    Settles B. Active Learning Literature Survey. 2009

    2009

  30. [38]

    Supervised clustering-algorithms and benefits

    Eick CF, Zeidat N, Zhao Z. Supervised clustering-algorithms and benefits. In: 16Th IEEE international conference on tools with artificial intelligence. 2004. p. 774–6

    2004

  31. [39]

    An algorithm to cluster documents based on relevance

    Desai M, Spink A. An algorithm to cluster documents based on relevance. Inf Process Manag. 2005 Sep;41(5):1035–49

    2005

  32. [40]

    Seeded Sequential LDA: A Semi-Supervised Algorithm for Topic-Specific Analysis of Sentences

    Watanabe K, Baturo A. Seeded Sequential LDA: A Semi-Supervised Algorithm for Topic-Specific Analysis of Sentences. Social Science Computer Review. 2024 Feb;42(1):224–48

    2024

  33. [41]

    Improving topic models with latent feature word representations

    Nguyen DQ, Billingsley R, Du L, Johnson M. Improving topic models with latent feature word representations. Transactions of the Association for Computational Linguistics. 2015;3:299–313

    2015

  34. [42]

    Using structural topic modeling to identify latent topics and trends in aviation incident reports

    Kuhn KD. Using structural topic modeling to identify latent topics and trends in aviation incident reports. Transp Res Part C Emerg Technol. 2018 Feb;87:105–22

    2018

  35. [43]

    Semisupervised fuzzy clustering with partition information of subsets

    Mei JP. Semisupervised fuzzy clustering with partition information of subsets. IEEE Transactions on Fuzzy Systems. 2018;27(9):1726–37

    2018

  36. [44]

    Self-organizing weighted incremental probabilistic latent semantic analysis

    Li N, Luo W, Yang K, Zhuang F, He Q, Shi Z. Self-organizing weighted incremental probabilistic latent semantic analysis. Int J Mach Learn Cybern. 2018 Dec;9(12):1987–98

    2018

  37. [45]

    Dataless text classification: A topic modeling approach with document manifold

    Li X, Li C, Chi J, Ouyang J, Li C. Dataless text classification: A topic modeling approach with document manifold. In: Proceedings of the 27th ACM international conference on information and knowledge management. 2018. p. 973–82

    2018

  38. [46]

    tBERT: Topic models and BERT joining forces for semantic similarity detection

    Peinelt N, Nguyen D, Liakata M. tBERT: Topic models and BERT joining forces for semantic similarity detection. In: Proceedings of the 58th annual meeting of the association for computational linguistics. 2020. p. 7047–55. Available from: https://aclanthology.org/2020.acl-main.630/ 34

    2020

  39. [47]

    Research on Multi-label Text Classification Method Based on tALBERT-CNN

    Liu W, Pang J, Li N, Zhou X, Yue F. Research on Multi-label Text Classification Method Based on tALBERT-CNN. Int J Comput Intell Syst. 2021 Dec;14(1):201

    2021

  40. [48]

    Supervised topic models for multi-label classification

    Li X, Ouyang J, Zhou X. Supervised topic models for multi-label classification. Neurocomputing. 2015;149:811–9

    2015

  41. [49]

    An overview of topic modeling and its current applications in bioinformatics

    Liu L, Tang L, Dong W, Yao S, Zhou W. An overview of topic modeling and its current applications in bioinformatics. SpringerPlus. 2016 Dec;5(1):1608

    2016

  42. [50]

    DOLDA: a regularized supervised topic model for high-dimensional multi-class regression

    Magnusson M, Jonsson L, Villani M. DOLDA: a regularized supervised topic model for high-dimensional multi-class regression. Comput Stat. 2020 Mar;35(1):175–201

    2020

  43. [51]

    Twin labeled LDA: a supervised topic model for document classification

    Wang W, Guo B, Shen Y, Yang H, Chen Y, Suo X. Twin labeled LDA: a supervised topic model for document classification. Appl Intell. 2020 Dec;50(12):4602–15

    2020

  44. [52]

    Effective document labeling with very few seed words: A topic model approach

    Li C, Xing J, Sun A, Ma Z. Effective document labeling with very few seed words: A topic model approach. In: Proceedings of the 25th ACM international on conference on information and knowledge management. 2016. p. 85–94

    2016

  45. [53]

    Labelset topic model for multi-label document classification

    Li X, Ouyang J, Zhou X. Labelset topic model for multi-label document classification. J Intell Inf Syst. 2016 Feb;46(1):83–97

    2016

  46. [54]

    Dataless Text Classification: A Topic Modeling Approach with Document Manifold

    Li X, Li C, Chi J, Ouyang J, Li C. Dataless Text Classification: A Topic Modeling Approach with Document Manifold. In: Proceedings of the 27th ACM International Conference on Information and Knowledge Management. Torino Italy: ACM; 2018. p. 973–82. Available from: https://dl.a...

    2018

  47. [55]

    Multi-label dataless text classification with topic modeling

    Zha D, Li C. Multi-label dataless text classification with topic modeling. Knowledge and Information Systems. 2019;61:137–60

    2019

  48. [56]

    Seed-guided deep document clustering

    Fard MM, Thonet T, Gaussier E. Seed-guided deep document clustering. In: Lecture Notes in Computer Science. Cham: Springer International Publishing; 2020. p. 3–16. (Lecture notes in computer science)

    2020

  49. [57]

    Community detection in social networks considering topic correlations

    Wang Y, Jin D, Musial K, Dang J. Community detection in social networks considering topic correlations. In: Proceedings of the AAAI Conference on Artificial Intelligence. 2019. p. 321–8. Available from: https://aaai.org/ojs/index.php/AAAI/article/view/3801

    2019

  50. [58]

    Transferable adversarial examples can efficiently fool topic models

    Wang Z, Zheng Y, Zhu H, Yang C, Chen T. Transferable adversarial examples can efficiently fool topic models. Computers & Security. 2022 Jul 1;118:102749

    2022

  51. [59]

    Topic extraction from extremely short texts with variational manifold regularization

    Li X, Wang Y, Ouyang J, Wang M. Topic extraction from extremely short texts with variational manifold regularization. Mach Learn. 2021 May;110(5):1029–66

    2021

  52. [60]

    HiGitClass: Keyword-driven hierarchical classification of GitHub repositories

    Zhang Y, Xu FF, Li S, Meng Y, Wang X, Li Q, et al. HiGitClass: Keyword-driven hierarchical classification of GitHub repositories. In: 2019 IEEE International Conference on Data Mining (ICDM). Beijing, China: IEEE; 2019

    2019

  53. [61]

    TherapyView: Visualizing Therapy Sessions with Temporal Topic Modeling and AI-Generated Arts

    Lin B, Zecevic S, Bouneffouf D, Cecchi G. TherapyView: Visualizing Therapy Sessions with Temporal Topic Modeling and AI-Generated Arts. arXiv; 2023. Available from: http://arxiv.org/abs/2302.10845

    2023 arXiv

  54. [62]

    Topic sentiment mixture: modeling facets and opinions in weblogs

    Mei Q, Ling X, Wondra M, Su H, Zhai C. Topic sentiment mixture: modeling facets and opinions in weblogs. In: Proceedings of the 16th international conference on World Wide Web. Banff Alberta Canada: ACM; 2007. p. 171–80. Available from: https://dl.acm.org/doi/10.1145/1242572.1242596

    2007

  55. [63]

    A study of classification of texts into categories of cybersecurity incident and attack with topic models

    Ishii M, Matsuura S, Mori K, Tomoishi M, Jin Y, Kitaguchi Y. A study of classification of texts into categories of cybersecurity incident and attack with topic models. In: Proceedings of the 6th International Conference on Information Systems Security and Privacy. Valletta, Ma...

    2020

  56. [65]

    A two-dimensional topic-Aspect Model for discovering multi-faceted topics

    Paul M, Girju R. A two-dimensional topic-Aspect Model for discovering multi-faceted topics. Proc Conf AAAI Artif Intell. 2010 Jul;24(1):545–50

    2010

  57. [2008]

    Available from: https://scienceopen.com/document?vid=6d552894-2cc3-4e2b-a483-41fa48a37ef8

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.