Pith. sign in

REVIEW 5 major objections 6 minor 1 cited by

SentiQNF: A Novel Approach to Sentiment Analysis Using Quantum Algorithms and Neuro-Fuzzy Systems

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A two-qubit 'quantum fuzzy' circuit is claimed to classify tweets with 100% accuracy.

desk verdict Internally inconsistent accuracy tables sink an otherwise competent survey of variational quantum classifiers for tweet sentiment. read the letter →

arxiv 2412.12731 v2 pith:DBAI46XC submitted 2024-12-17 cs.CL quant-ph

Kshitij Dave , Nouhaila Innan , Bikash K. Behera , Zahid Mumtaz , Saif Al-Kuwari , Ahmed Farouk This is my paper · ORCID
classification cs.CLquant-ph
keywords quantumfuzzyneuralnetworksentimentanalysisvariationalclassifierneuro-fuzzysystemshybridquantum-classicalmodelsnoiserobustnessbinarytweetclassificationTF-IDFencoding
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

The paper proposes a hybrid sentiment classifier called the Quantum Fuzzy Neural Network (QFNN): a small variational quantum circuit whose extra rotation layers are described as a fuzzy membership layer, trained with classical optimization. Its central claim is that QFNN outperforms every classical, quantum, and hybrid baseline tested, reaching 100% accuracy on a two-class subset of the Coronavirus Tweets Dataset and 90% on a two-class subset of the General Sentimental Tweets Dataset. The authors also claim the model stays accurate under six quantum noise models, positioning it as a candidate for noisy, large-scale sentiment processing. If the claim held, it would mean a two-qubit classifier with a small set of trainable rotations can beat much larger classical models, which is why the result depends heavily on how the binary tweet subsets were built.

What carries the argument

The load-bearing object is a two-qubit variational circuit: TF-IDF word features are angle-embedded into two qubits, alternating RX and RY rotations with a controlled-Z gate create entanglement, and an extended 'fuzzy' block of additional RX and RY rotations acts as the membership layer before measurement. Eight parameters $\theta_1,\ldots,\theta_8$ are trained with the ADAM optimizer against mean squared error. The fuzzy layer is realized as rotation blocks in the quantum circuit rather than as a separate classical module, and the paper argues these rotations encode degrees of word-to-sentiment association.

What would settle it

Train a simple logistic regression on the exact same 1000/500 binary splits with the same TF-IDF features and a fixed random seed; if it matches or beats QFNN, the reported 100% and 90% accuracies are not evidence for a quantum or fuzzy advantage.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that adding an extended block of parameterized RX and RY rotations, called the fuzzy composition layer, between angle embedding and measurement lets a two-qubit QFNN achieve 100% accuracy on the binary CVTD subset and 90% on the binary GSTD subset, while the classical baselines peak below 82% on CVTD and below 80% on GSTD. The authors present this as evidence that quantum superposition combined with fuzzy membership captures linguistic uncertainty that classical pipelines miss. They further report that the model maintains accuracy under depolarizing noise even at high noise levels, whereas other noise channels such as bit flip and phase flip can drive accuracy to zero in certain ranges.

Load-bearing premise

The result rests on how the two-class tweet subsets were built from the original five-class sets; the paper does not report the feature-selection rule or the random seed, so if those subsets were easy or contaminated by test information, the reported accuracies would not reflect a genuine quantum or fuzzy advantage.

Editorial extensions

If this is right

  • If the reported accuracies are right, the same two-qubit circuit with eight trainable parameters outperforms all tested classical, quantum, and hybrid classifiers on both binary tweet subsets.
  • The robustness results imply the model can operate under depolarizing noise without dropping to zero accuracy, suggesting noisy near-term hardware could run it without full error correction.
  • The reported success of the fuzzy rotation layer implies that adding fuzziness inside a quantum circuit, rather than only to classical networks, is a viable direction for sentiment classification.
  • Because the gains appear on binary subsets with neutral tweets removed, the claim is specifically about separating clear positive from clear negative tweets, not about classifying the original five sentiment labels.
  • The comparison table against prior quantum sentiment models places QFNN as the top-scoring quantum or hybrid approach on these datasets.

Reading between the lines

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

  • A direct test of the claimed advantage is to remove the controlled-Z entanglement gate while keeping the fuzzy rotation block; if accuracy barely changes, the entanglement is not load-bearing.
  • The natural next experiment is the same architecture on the original five-class sentiment labels, which the authors explicitly list as a current limitation.
  • A classical network with the same number of trainable parameters and the same TF-IDF features would isolate whether the gain comes from the variational circuit or simply from fitting a small, easy binary subset.
  • The absence of a reported random seed and feature-selection rule means the headline accuracies should be treated as specific to the paper's exact binary pipeline until reproduced independently.
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

5 major / 6 minor

Summary. The paper proposes SentiQNF, a “Quantum Fuzzy Neural Network” (QFNN) for sentiment analysis, in which TF-IDF features are angle-embedded into a two- or four-qubit circuit and a set of parameterized Rx/Ry rotations is described as a fuzzy layer. The authors report experiments on two Twitter datasets (CVTD and GSTD), claim that QFNN achieves 100% and 90% accuracy respectively and thereby outperforms all classical, quantum, and hybrid baselines, and further claim robustness under six noise models. The manuscript also contains three theoretical propositions intended to justify a quantum advantage for sentiment analysis.

Significance. If the central claims were valid, a two-qubit variational classifier with a renamed parameterized layer beating all classical and quantum baselines on sentiment analysis would be a surprising result worth careful scrutiny. The paper does provide pseudocode for QFNN, HQNN, and HFNN, a comparative accuracy table, and a broad set of baselines, which are useful presentation features. However, the headline empirical numbers are internally inconsistent, the theoretical propositions do not establish the claimed advantages, and the data pipeline is not reproducible as described. The manuscript therefore does not support its central claim. No code, data splits, or random seeds are supplied, so the experiments cannot be independently checked.

major comments (5)
  1. [Section IV-E, Tables III and V; Abstract; Section V] The reported QFNN accuracy is logically inconsistent with the error rates in the same tables. In Table III, QFNN on CVTD is listed with 100% accuracy but also FP rate = 10.81%, FN rate = 27.09%, and FD rate = 12.98%; 100% binary accuracy implies zero misclassifications and therefore zero false positives, false negatives, and false discoveries. Table V is similarly impossible: QFNN on GSTD is listed with 90% accuracy but FN rate = 37.33% and FD rate = 29.64%, which would require more than the 10% error budget unless the rates are computed over different subsets or with nonstandard definitions, and no such definitions are provided. Since the Abstract and Section V base the “outperforms all” conclusion directly on these numbers, the central empirical result is invalid as reported.
  2. [Section III-B, Propositions 1–3] The theoretical propositions do not prove the claims they are used for. Proposition 1 derives that two complex amplitudes with the same squared modulus have the same modulus, which is a trivial restatement rather than a demonstration that “quantum probability’s versatility captures linguistic uncertainty.” Proposition 2’s nonlinearity argument appeals to an undefined homogeneity condition kp(x) = p(kx) and a contradiction “at x = π/2” where x is not clearly a word, angle, or probability, and the derivation of p(c) in Eq. (19) omits normalization and uses α and β ambiguously. Proposition 3 is a description of tensor-product composition, not a proof that the composite system “reflects the relations between individual elements and the entire context.” These propositions are the only theoretical justification for a quantum-specific advantage, and they do not support it.
  3. [Sections IV-A, IV-B, IV-C and Table I] The empirical pipeline is not reproducible from the manuscript. The five-class CVTD and the multi-class GSTD are collapsed to binary by excluding neutral sentiments, but the paper does not report the resulting class sizes or balance, the number of TF-IDF features after selection, the feature-selection rule that maps those features onto 2 or 4 qubits, the random seed for the 1000/500 split, or the final hyperparameters actually used for the reported QFNN result. Table I states test splits of 20%, 30%, and 50% while Sections IV-C and IV-E state a fixed 1000/500 split. Without these details, the reported “outperforms all” claim may be an artifact of the specific constructed binary subset rather than of any quantum or fuzzy processing, and the experiment cannot be replicated.
  4. [Sections III-D, III-E, III-F and Algorithm 1] The “fuzzy layer” is not an implementation of fuzzy logic. In Eq. (21), the membership function μ(w, S_i) is defined as a weighted sum over words with weights w_j, but those weights are never specified or learned, and the association function F(w_j, S_i) is left undefined. In Eq. (27), the “modified projection operator” P'_i is a sum of scalar multiples of the same projector |S_i⟩⟨S_i|, which collapses to a scalar times the projector and does not create the claimed fuzzy measurement. The circuit in Fig. 2 and Eq. (31) is a standard parameterized variational layer of Rx/Ry/CZ gates; calling it “fuzzy” is a naming choice. Consequently, the comparison with CF and HFNN does not test a quantum-fuzzy hybrid in any established sense, undermining the title and the claimed novelty.
  5. [Section IV-F and Fig. 7] The noise-robustness claim is not supported by the evidence shown. Fig. 7 reports QFNN accuracy under six noise models without error bars, repeated-seed statistics, or a noiseless baseline for comparison, so the reader cannot judge whether fluctuations are significant. The text also appears to contradict the figure: it says that for CVTD the depolarizing model “maintains non-zero accuracy even at the highest noise levels” while the other models drop to 0%, but the figure legend and surrounding discussion are not quantitative enough to verify this. Since robustness is one of the three research questions in Section I-C, this part needs a reproducible protocol with seeds, repetitions, and a clear definition of the noise parameter range.
minor comments (6)
  1. [Eq. (12)] The notation in Eq. (12) is inconsistent: the sample index is written as γ but the left-hand side uses X_i, and the subscripts α, β, γ, t, ϕ are not defined as ranging over any specific sets. Please rewrite the sample notation and define all subscripts.
  2. [Algorithm 1, line 11] Algorithm 1 initializes random QFNN parameters {theta_1, ..., theta_8} inside the per-batch loop, which would reset the parameters at every batch and prevent actual gradient-based training; the initialization should occur before the epoch loop. This may also explain the unusual loss/accuracy curves in Figs. 4 and 5.
  3. [Eq. (31) and Table I] The text says the fuzzy layer uses parameters theta_5 through theta_8, but Eq. (31) applies theta_5 on both qubits, then theta_6, theta_7, and theta_8 twice each, which matches neither the eight-parameter description in Algorithm 1 nor the 16/19-parameter counts in Table I. Please make the layer definitions and parameter counts consistent.
  4. [Section IV-B] There is a duplicated phrase: “systematically starts by systematically preprocessing”; this should be corrected to a single occurrence.
  5. [Section V and Fig. 7] The text refers to the “AP error model”, but only six noise models (BF, PF, BPF, DP, AD, PD) are defined in Section IV-D; this is likely a typo for AD or PD and should be fixed.
  6. [Page 12 (end matter)] The final part of the manuscript contains garbled “/uni...” sequences in the text flow, apparently a font or encoding corruption; these should be removed or repaired before any resubmission.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported accuracies are empirical results from a held-out split, not quantities derived by construction from fitted inputs or self-citations.

full rationale

The reported QFNN accuracies (100% on CVTD, 90% on GSTD) are empirical results obtained from a 1000/500 train/test split, not quantities derived by construction from the fitted parameters or from the definitions of the fuzzy layer. The only author self-citation touching the method is [33], used to name standard Kraus noise channels; the robustness curves in Figs. 7a and 7b are generated by the present simulations, and the noise-model definitions are parameter-free standard channels, so this citation is not load-bearing. The 'fuzzy layer' is a sequence of RX/RY and CZ rotations (Eq. 31 and Algorithm 1), and calling it fuzzy is a labeling or construct-validity choice rather than a circular step: no equation in the paper defines the target accuracy in terms of the fuzzy label. Likewise, the choice of EfficientSU2 over RealAmplitude is a model-selection report, not a fitted input renamed as a prediction. The internal contradiction between 100% accuracy and nonzero FP/FN/FD rates in Table III, and the failure to report a seed or exact binary subset, are serious correctness and reproducibility concerns, but they are not circularity. I therefore find no significant circularity and score 0.

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

No new physical entities are introduced; the paper renames a parameterized layer as 'fuzzy', which is an architectural label rather than a new postulated entity. The load-bearing extras are free parameters and domain assumptions listed above.

free parameters (5)
  • QFNN variational angles theta_1 through theta_8 = 8 values, not reported
    All reported accuracy depends on these parameters trained with ADAM on the training split; they are fitted, not derived.
  • Feature selection rule for TF-IDF features = unspecified
    The pipeline says features are selected but gives no count, threshold, or algorithm, and the 2-qubit circuit needs a specific reduction; this controls the input.
  • Train/test split and neutral-class exclusion = 1000 train, 500 test; seed unspecified
    Section IV-C fixes dataset size, but not how rows are sampled or whether selection happens before feature reduction; this can decide the headline accuracy.
  • Hyperparameters (learning rate, epochs, ansatz type) = LR 0.1/0.01/0.001; epochs 20/30/100; EfficientSU2
    Section IV-C reports choosing EfficientSU2 because it outperformed RealAmplitude, without a separate validation set, so the choice is tuned to the test data.
  • Fuzzy membership weights w_j in Equation 21 = unspecified
    The fuzzy measurement uses weights assigned by importance but no method is given, so the fuzzy layer's behavior is not fully specified.
assumptions (5)
  • domain assumption TF-IDF term statistics are sufficient sentiment features.
    Equations 13-14 and Section IV-B define the entire input representation; no pretrained embeddings or context model is used.
  • domain assumption Neutral removal and five-to-two class collapse preserves the sentiment task.
    Section IV-A and IV-B explicitly drop neutral tweets and convert five sentiment classes into binary labels; this directly makes the reported 100% accuracy easier.
  • domain assumption Qiskit qasm simulator and Kraus noise channels faithfully model device noise.
    Section IV-D and IV-E evaluate robustness only on a simulated backend, so noise robustness is not demonstrated on quantum hardware.
  • ad hoc to paper Parameterized Rx and Ry rotations constitute a fuzzy layer.
    Section III-D and Algorithm 1 describe the 'fuzzy layer' only as additional rotation gates; no fuzzy rule base, membership inference, or defuzzification step is implemented.
  • ad hoc to paper Quantum superposition gives sentiment analysis an inherent advantage.
    Propositions 1-3 in Section III-B are incomplete and non-sequitur arguments that are not connected to the measured circuit performance.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SentiQNF: A Novel Approach to Sentiment Analysis Using Quantum Algorithms and Neuro-Fuzzy Systems." pith.science (2026). https://pith.science/paper/DBAI46XC

@misc{pith2026241212731,
  author       = {Pith},
  title        = {Pith review of: SentiQNF: A Novel Approach to Sentiment Analysis Using Quantum Algorithms and Neuro-Fuzzy Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DBAI46XC}},
  note         = {Machine review of arXiv:2412.12731}
}
read the original abstract

Sentiment analysis is an essential component of natural language processing, used to analyze sentiments, attitudes, and emotional tones in various contexts. It provides valuable insights into public opinion, customer feedback, and user experiences. Researchers have developed various classical machine learning and neuro-fuzzy approaches to address the exponential growth of data and the complexity of language structures in sentiment analysis. However, these approaches often fail to determine the optimal number of clusters, interpret results accurately, handle noise or outliers efficiently, and scale effectively to high-dimensional data. Additionally, they are frequently insensitive to input variations. In this paper, we propose a novel hybrid approach for sentiment analysis called the Quantum Fuzzy Neural Network (QFNN), which leverages quantum properties and incorporates a fuzzy layer to overcome the limitations of classical sentiment analysis algorithms. In this study, we test the proposed approach on two Twitter datasets: the Coronavirus Tweets Dataset (CVTD) and the General Sentimental Tweets Dataset (GSTD), and compare it with classical and hybrid algorithms. The results demonstrate that QFNN outperforms all classical, quantum, and hybrid algorithms, achieving 100% and 90% accuracy in the case of CVTD and GSTD, respectively. Furthermore, QFNN demonstrates its robustness against six different noise models, providing the potential to tackle the computational complexity associated with sentiment analysis on a large scale in a noisy environment. The proposed approach expedites sentiment data processing and precisely analyses different forms of textual data, thereby enhancing sentiment classification and insights associated with sentiment analysis.

Figures

Figures reproduced from arXiv: 2412.12731 by the authors.

Figure 1
Figure 1. The architecture of (a) classical fuzzy (CF) model. (b) QFNN model. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The QFNN Circuit. These gates characterized controlled rotations on the Bloch sphere, enabling the network to manipulate qubit states and represent fuzzy membership values with enhanced precision. The controlled-Rx and controlled-Ry gates can be described as, CRx(θ) = |0⟩⟨0| ⊗ I + |1⟩⟨1| ⊗ Rx(θ), (24) CRy(θ) = |0⟩⟨0| ⊗ I + |1⟩⟨1| ⊗ Ry(θ). (25) Combining fuzzy logic, Rx, and Ry gates in the fuzzy composition layer em… view at source ↗
Figure 3
Figure 3. (a) Flowchart illustrating the preprocessing steps for sentiment analysis, and (b) represents an example of preprocessing steps applied [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: The loss and accuracy of GSTD for (a) ANN, (b) QNN, (c) [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 7
Figure 7. Figure 7: The accuracy of the QFNN in the presence of six noise models [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

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. Visualizing Public Opinion on X: A Real-Time Sentiment Dashboard Using VADER and DistilBERT

    econ.GN 2025-04 reject novelty 3.0 of 10

    A VADER-plus-DistilBERT ensemble is reported with 87.6% accuracy on a manually labeled tweet set, but the evaluation set is the same set used to tune the ensemble weight, and no artifacts are released.

Reference graph

Works this paper leans on

36 extracted references · 26 canonical work pages · cited by 1 Pith paper

  1. [1]

    Quantum natural language processing based sentiment analysis using lambeq toolkit,

    S. Ganguly, S. N. Morapakula, and L. M. P. Coronado, “Quantum natural language processing based sentiment analysis using lambeq toolkit,” in 2022 Second International Conference on Power, Control and Computing Technologies (ICPC2T). IEEE, 2022, pp. 1–6. [Online]. Available: https://doi.org/10.1109/ICPC2T53885.2022.9776836

    work page arXiv 2022

  2. [2]

    A survey on sentiment analysis by using machine learning methods,

    P. Yang and Y . Chen, “A survey on sentiment analysis by using machine learning methods,” in 2017 IEEE 2nd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC). IEEE, 2017, pp. 117–121. [Online]. Available: https: //doi.org/10.1109/ITNEC.2017.8284920

    work page arXiv 2017

  3. [3]

    Optimization of sentiment analysis using machine learning classifiers,

    J. Singh, G. Singh, and R. Singh, “Optimization of sentiment analysis using machine learning classifiers,” Human-centric Computing and information Sciences , vol. 7, pp. 1–12, 2017. [Online]. Available: https://doi.org/10.1186/s13673-017-0116-3

    doi:10.1186/s13673-017-0116-3 2017

  4. [4]

    Sentiment analysis on product reviews using machine learning techniques,

    R. S. Jagdale, V . S. Shirsat, and S. N. Deshmukh, “Sentiment analysis on product reviews using machine learning techniques,” in Cognitive Informatics and Soft Computing: Proceeding of CISC

    work page

  5. [5]

    Sentiment analysis of movie reviews using machine learning techniques,

    S. Kumar, K. Sharma, D. Veragi, and A. Juyal, “Sentiment analysis of movie reviews using machine learning techniques,” 2022 International Conference on Machine Learning, Big Data, Cloud and Parallel Computing (COM-IT-CON) , 2022. [Online]. Available: https://doi.org/10.1109/COM-IT-CON54601.2022.9850878

    arXiv 2022

  6. [6]

    Application of machine learning techniques to sentiment analysis,

    A. P. Jain and P. Dandannavar, “Application of machine learning techniques to sentiment analysis,” in 2016 2nd International Conference on Applied and Theoretical Computing and Communication Technology (iCATccT). IEEE, 2016, pp. 628–632. [Online]. Available: https: //doi.org/10.1109/ICATCCT.2016.7912076

    arXiv 2016

  7. [7]

    A review of contextual information for context-based approach in sentiment analysis,

    N. N. Yusof, A. Mohamed, and S. Abdul-Rahman, “A review of contextual information for context-based approach in sentiment analysis,” International Journal of Machine Learning and Computing , vol. 8, no. 4, pp. 399–403, 2018. [Online]. Available: https: //doi.org/10.18178/ijmlc.2018.8.4.719

    work page doi:10.18178/ijmlc.2018.8.4.719 2018

  8. [8]

    Who cares about sarcastic tweets? investigating the impact of sarcasm on sentiment analysis,

    D. G. Maynard and M. A. Greenwood, “Who cares about sarcastic tweets? investigating the impact of sarcasm on sentiment analysis,” in Lrec 2014 proceedings . ELRA, 2014. [Online]. Available: https://eprints.whiterose.ac.uk/130763/

    work page 2014

Show all 36 references
  1. [9]

    Enhancement bag-of-words model for solving the challenges of sentiment analysis,

    D. M. El-Din, “Enhancement bag-of-words model for solving the challenges of sentiment analysis,” International Journal of Advanced Computer Science and Applications , vol. 7, no. 1, 2016. [Online]. Available: https://doi.org/10.14569/IJACSA.2016.070134

    2016

  2. [10]

    Sentiment analysis of extremism in social media from textual information,

    M. Asif, A. Ishtiaq, H. Ahmad, H. Aljuaid, and J. Shah, “Sentiment analysis of extremism in social media from textual information,” Telematics and Informatics , vol. 48, p. 101345, 2020. [Online]. Available: https://doi.org/10.1016/j.tele.2020.101345

    2020

  3. [11]

    The biases of pre-trained language models: An empirical study on prompt-based sentiment analysis and emotion detection,

    R. Mao, Q. Liu, K. He, W. Li, and E. Cambria, “The biases of pre-trained language models: An empirical study on prompt-based sentiment analysis and emotion detection,” IEEE Transactions on Affective Computing, 2022. [Online]. Available: https://doi.org/10.1109/ TAFFC.2022.3204972

    2022

  4. [12]

    Sentiment analysis in twitter using machine learning techniques,

    M. Neethu and R. Rajasree, “Sentiment analysis in twitter using machine learning techniques,” in 2013 fourth international conference on computing, communications and networking technologies (ICCCNT) . IEEE, 2013, pp. 1–5. [Online]. Available: https://doi.org/10.1109/ ICCCNT.2...

    2013

  5. [13]

    Rapid solution of problems by quantum computation,

    D. Deutsch and R. Jozsa, “Rapid solution of problems by quantum computation,” Proceedings of the Royal Society of London. Series A: Mathematical and Physical Sciences , vol. 439, no. 1907, pp. 553–558,

    1907

  6. [14]

    Quantum machine learning,

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, “Quantum machine learning,” Nature, vol. 549, no. 7671, pp. 195–202, 2017. [Online]. Available: https://doi.org/10.1038/nature23474

    2017 doi

  7. [15]

    Deep sentiment extraction using fuzzy-rule based deep sentiment analysis,

    S. JASTI and G. V . S. R. KUMAR, “Deep sentiment extraction using fuzzy-rule based deep sentiment analysis,” International Journal of Advanced Computer Science and Applications , vol. 13, no. 6, 2022. [Online]. Available: https://dx.doi.org/10.14569/IJACSA.2022.0130650

    2022

  8. [16]

    Sentiment analysis of students’ feedback on e-learning using a hybrid fuzzy model,

    M. Alzaid and F. Fkih, “Sentiment analysis of students’ feedback on e-learning using a hybrid fuzzy model,” Appl. Sci. , vol. 13, no. 23, p. 12956, 2023. [Online]. Available: https://www.mdpi.com/2076-3417/ 13/23/12956

    2023

  9. [17]

    Fuzzy sentiment analysis using convolutional neural network,

    Sugiyarto, J. Eliyanto, N. Irsalinda, and M. Fitrianawati, “Fuzzy sentiment analysis using convolutional neural network,” AIP Conf. Proc., vol. 2329, p. 050002, 2021. [Online]. Available: https: //doi.org/10.1063/5.0042144

    2021 doi

  10. [18]

    Fuzzy deep belief networks for semi-supervised sentiment classification,

    S. Zhou, Q. Chen, and X. Wang, “Fuzzy deep belief networks for semi-supervised sentiment classification,” Neurocomputing, vol. 131, pp. 312–322, 2014. [Online]. Available: https://doi.org/10.1016/j. neucom.2013.10.011

    2014 doi

  11. [19]

    Sentiment analysis on twitter data using modified neuro fuzzy algorithm,

    P. N. S. and S. Karthik, “Sentiment analysis on twitter data using modified neuro fuzzy algorithm,” Int. J. Intel. Sys. Appl. Eng., vol. 11, no. 10S, p. 370–384, 2023. [Online]. Available: https://ijisae.org/index.php/IJISAE/article/view/3292

    2023

  12. [20]

    A survey of quantum-cognitively inspired sentiment analysis models,

    Y . Liu, Q. Li, B. Wang, Y . Zhang, and D. Song, “A survey of quantum-cognitively inspired sentiment analysis models,” ACM Computing Surveys, vol. 56, no. 1, pp. 1–37, 2023. [Online]. Available: https://doi.org/10.1145/3604550

    2023 doi

  13. [21]

    Quantum-inspired fully complex-valued neutral network for sentiment analysis,

    W. Lai, J. Shi, and Y . Chang, “Quantum-inspired fully complex-valued neutral network for sentiment analysis,” Axioms, vol. 12, no. 3, p. 308,

  14. [22]

    The role of entanglement for enhancing the efficiency of quantum kernels towards classification,

    D. Sharma, P. Singh, and A. Kumar, “The role of entanglement for enhancing the efficiency of quantum kernels towards classification,” Physica A: Statistical Mechanics and its Applications , vol. 625, p. 128938, 2023. [Online]. Available: https://doi.org/10.1016/j.physa.2023. 128938

    2023 doi

  15. [23]

    Quantum representation for sentiment classification,

    F. Z. Ruskanda, M. R. Abiwardani, M. A. Al Bari, K. A. Bagaspati, R. Mulyawan, I. Syafalni, and H. T. Larasati, “Quantum representation for sentiment classification,” in 2022 IEEE International Conference on Quantum Computing and Engineering (QCE) . IEEE, 2022, pp. 67–78. [Onl...

    2022

  16. [24]

    Quantum-enhanced support vector machine for sentiment classification,

    F. Z. Ruskanda, M. R. Abiwardani, R. Mulyawan, I. Syafalni, and H. T. Larasati, “Quantum-enhanced support vector machine for sentiment classification,” IEEE Access , 2023. [Online]. Available: https://doi.org/10.1109/ACCESS.2023.3304990

    2023

  17. [25]

    Q-lstm language model–decentralized quantum multilingual pre-trained language model for privacy protection,

    S. S. Li, X. Zhang, S. Zhou, H. Shu, R. Liang, H. Liu, and L. P. Garcia, “Q-lstm language model–decentralized quantum multilingual pre-trained language model for privacy protection,” arXiv preprint arXiv:2210.03221 , 2022. [Online]. Available: https: //doi.org/10.48550/arXiv.2...

    2022 doi arXiv

  18. [26]

    Fuzzy logic,

    L. A. Zadeh, “Fuzzy logic,” Computer, vol. 21, no. 4, pp. 83–93, 1988. [Online]. Available: https://doi.org/10.1109/2.53

    1988 doi

  19. [27]

    Chapter 2 - fundamentals of type-1 fuzzy logic theory,

    E. Kayacan and M. A. Khanesar, “Chapter 2 - fundamentals of type-1 fuzzy logic theory,” in Fuzzy Neural Networks for Real Time Control Applications , E. Kayacan and M. A. Khanesar, Eds. Butterworth-Heinemann, 2016, pp. 13–24. [Online]. Available: https://doi.org/10.1016/B978-0...

    2016 doi

  20. [28]

    Coronavirus tweets nlp - text classification

    “Coronavirus tweets nlp - text classification.” [Online]. Available: https: //www.kaggle.com/datasets/datatattle/covid-19-nlp-text-classification 13

  21. [29]

    Sentiment140 dataset with 1.6 million tweets

    “Sentiment140 dataset with 1.6 million tweets.” [Online]. Available: https://www.kaggle.com/datasets/kazanova/sentiment140

  22. [30]

    Hybrid quantum-classical machine learning for sentiment analysis,

    A. K. M. Masum, A. Maurya, D. S. Murthy, Pratibha, and N. Mahmud, “Hybrid quantum-classical machine learning for sentiment analysis,” arXiv:2310.10672, 2023. [Online]. Available: https://doi.org/10.48550/ arXiv.2310.10672

    2023 doi arXiv

  23. [31]

    Quantum computing and machine learning for arabic language sentiment classification in social media,

    A. Omar and T. A. El-Hafeez, “Quantum computing and machine learning for arabic language sentiment classification in social media,” Scientific Reports , vol. 13, p. 17305, 2023. [Online]. Available: https://www.nature.com/articles/s41598-023-44113-7

    2023

  24. [32]

    Cfn: A complex-valued fuzzy network for sarcasm detection in conversations,

    Y . Zhang, Y . Liu, Q. Li, P. Tiwari, B. Wang, Y . Li, H. M. Pandey, P. Zhang, and D. Song, “Cfn: A complex-valued fuzzy network for sarcasm detection in conversations,” IEEE TRANSACTIONS ON FUZZY SYSTEMS, vol. 29, no. 12, p. 3696, 2021. [Online]. Available: https://doi.org/10...

    2021

  25. [33]

    Analysis of quantum machine learning algorithms in noisy channels for classification tasks in the iot extreme environment,

    S. K. Satpathy, V . Vibhu, B. K. Behera, S. Al-Kuwari, S. Mumtaz, and A. Farouk, “Analysis of quantum machine learning algorithms in noisy channels for classification tasks in the iot extreme environment,” IEEE Internet of Things Journal , pp. 1 – 1, 2023. [Online]. Available:...

    2023

  26. [1992]

    Available: https://doi.org/10.1098/rspa.1992.0167

    [Online]. Available: https://doi.org/10.1098/rspa.1992.0167

    1992

  27. [2017]

    Springer, 2019, pp. 639–647. [Online]. Available: https: //doi.org/10.1007/978-981-13-0617-4 61

    2019 doi

  28. [2023]

    Available: https://doi.org/10.3390/axioms12030308

    [Online]. Available: https://doi.org/10.3390/axioms12030308

    doi

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.