Multichannel Two-Dimensional Convolutional Neural Network Based on Interactive Features and Group Strategy for Chinese Sentiment Analysis
Abstract
:1. Introduction
- The character-based integer encoding method is applied to the text sequence. Through this method, the feature extraction network can flexibly extract different levels of feature information, which is beneficial to retain more fine-grained information.
- The interactive features between character vector elements corresponding to any sample are introduced to improve the dimensionality of feature vectors and enhance sentiment features. We can obtain the corresponding semantic information by controlling the elements involved in the interaction.
- Group strategy has successfully realized the conversion of feature sequences into feature maps. Compared with traditional feature sequences, feature maps contain more feature information so that more sentiment features can be learned by the model.
- Multichannel two-dimensional convolutional neural networks with different convolution kernel sizes are used to extract sentiment features of different scales, which can effectively avoid the problem that single-channel networks cannot fully extract sentiment features.
- In the existing literature on sentiment analysis methods, we have implemented sentiment analysis based on two-dimensional convolutional neural networks for the first time and proposed a series of methods to ensure that two-dimensional convolutional neural networks are successfully applied to sentiment analysis tasks.
2. Related Work
3. Our Method
3.1. Character-Level Encoding
3.2. Interactive Features
3.2.1. Definition of Interaction Features
3.2.2. Representation of Interactive Feature Vectors
- represents the vector continuous concatenation symbol. Assuming , then:
- represents the combining marker in mathematics
- has the same effect as
- means to fuse two vectors into a two-dimensional matrix, for example:
- represents the first term of the interactive feature vector
- represents the interactive feature vector after removing the first interactive features in
- represents the interactive feature vector generated by element
- where , , , the initial interactive feature vector .
3.2.3. The Determination Principle of
3.3. Group Strategy
3.4. Multichannel Two-Dimensional Convolutional Neural Networks
4. Experiments
4.1. Experiment Setups
4.1.1. Datasets
4.1.2. Data Preprocessing
4.1.3. Evaluation Metrics
4.1.4. Hyperparameter Setting
4.2. Comparison with Existing Methods
4.2.1. Comparative Methods
- LSTM: In this method, a LSTM network is used to extract sentiment features. It is composed of a embedding layer, an LSTM layer and a full connected layer [53].
- Two-layer LSTM: A two-layer LSTM is used to extract text features [54].
- BiLSTM: Bidirectional long short-term memory network. A one-layer BiLSTM is used to extract text features [54].
- Two-layer BiLSTM: A two-layer BiLSTM is used to extract text features [54].
- GRU: A gated recurrent unit is a variant of LSTM. Compared with LSTM, GRU retains its resistance to gradient disappearance. Meanwhile, its internal structure is simpler, training is faster, and it has been widely used for sentiment analysis recently [55].
- BiGRU: Bidirectional gated recurrent unit. In order to solve the difficult problem of modeling sentiment relationships in recurrent structure, Chen et al. [56] proposed to use a bidirectional gated recurrent unit to capture sentimental relationship information.
- Character-level ConvNets: This method applies ConvNets to characters for the first time. Experimental results show that when training on large-scale datasets, ConvNets do not require word-level information. In addition, existing research results show that ConvNets do not rely on semantic information [57].
- SLCABG: This method uses a convolutional neural network (CNN) combined with a bidirectional gated recurrent unit (BiGRU) for sentiment analysis of Chinese text. The attention mechanism and the sentiment lexicon are used to emphasize important information and enhance sentiment features, respectively [58].
- MDMLSM: This model firstly uses the pre-trained BERT model to form word vectors, then applies the attention-based BiLSTM to extract text features, and finally the output feature representations are sequentially input into the multilayer perceptron and sentiment classifier [59].
- MC-2D-CNN (word-based): In this article, the MCNN-IFGS we propose is based on character-based operation objects. In order to further explore the differences between characters and words, we added the word-based MCNN-IFGS as a comparison method.
4.2.2. Results and Analysis
4.3. Further Analysis of MCNN-IFGS
4.3.1. Effect of Learning Rate
4.3.2. Effect of Model Parameters
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Wang, L.; Niu, J.W.; Yu, S. SentiDiff: Combining Textual Information and Sentiment Diffusion Patterns for Twitter Sentiment Analysis. IEEE Trans. Knowl. Data Eng. 2020, 32, 2026–2039. [Google Scholar] [CrossRef]
- Hassonah, M.A.; Al-Sayyed, R.; Rodan, A.; Al-Zoubi, A.M.; Aljarah, I.; Faris, H. An efficient hybrid filter and evolutionary wrapper approach for sentiment analysis of various topics on Twitter. Knowl.-Based Syst. 2020, 192, 105353. [Google Scholar] [CrossRef]
- Martin-Gutierrez, D.; Hernandez-Penaloza, G.; Hernandez, A.B.; Lozano-Diez, A.; Alvarez, F. A Deep Learning Approach for Robust Detection of Bots in Twitter Using Transformers. IEEE Access 2021, 9, 54591–54601. [Google Scholar] [CrossRef]
- Xu, Y.Y.; Liu, Z.X.; Zhao, J.C.; Su, C.W. Weibo sentiments and stock return: A time-frequency view. PLoS ONE 2017, 12, e0180723. [Google Scholar] [CrossRef] [Green Version]
- Chen, J.Y.; Becken, S.; Stantic, B. Lexicon Based Chinese Language Sentiment Analysis Method. Comput. Sci. Inf. Syst. 2019, 16, 639–655. [Google Scholar] [CrossRef] [Green Version]
- Ling, M.J.; Chen, Q.H.; Sun, Q.; Jia, Y.B. Hybrid Neural Network for Sina Weibo Sentiment Analysis. IEEE Trans. Comput. Soc. Syst. 2020, 7, 983–990. [Google Scholar] [CrossRef]
- Kagan, V.; Stevens, A.; Subrahmanian, V.S. Using Twitter Sentiment to Forecast the 2013 Pakistani Election and the 2014 Indian Election. IEEE Intell. Syst. 2015, 30, 2–5. [Google Scholar] [CrossRef]
- Ibrahim, M.; Abdillah, O.; Wicaksono, A.F.; Adriani, M. Buzzer Detection and Sentiment Analysis for Predicting Presidential Election Results in A Twitter Nation. In Proceedings of the 2015 IEEE International Conference on Data Mining Workshop (Icdmw), Atlantic City, NJ, USA, 14–17 November 2015; pp. 1348–1353. [Google Scholar] [CrossRef]
- Baccianella, S.; Esuli, A.; Sebastiani, F. SENTIWORDNET 3.0: An Enhanced Lexical Resource for Sentiment Analysis and Opinion Mining. In Proceedings of the Lrec 2010-Seventh International Conference on Language Resources and Evaluation, Valletta, Malta, 19–21 May 2010. [Google Scholar]
- Deng, S.Y.; Sinha, A.P.; Zhao, H.M. Adapting sentiment lexicons to domain-specific social media texts. Decis. Support Syst. 2017, 94, 65–76. [Google Scholar] [CrossRef]
- Asghar, M.Z.; Khan, A.; Ahmad, S.; Qasim, M.; Khan, I.A. Lexicon-enhanced sentiment analysis framework using rule-based classification scheme. PLoS ONE 2017, 12, e0171649. [Google Scholar] [CrossRef]
- Han, H.Y.; Zhang, J.P.; Yang, J.; Shen, Y.R.; Zhang, Y.S. Generate domain-specific sentiment lexicon for review sentiment analysis. Multimed. Tools Appl. 2018, 77, 21265–21280. [Google Scholar] [CrossRef]
- Cai, Y.; Yang, K.; Huang, D.P.; Zhou, Z.K.; Lei, X.; Xie, H.R.; Wong, T.L. A hybrid model for opinion mining based on domain sentiment dictionary. Int. J. Mach. Learn. Cybern. 2019, 10, 2131–2142. [Google Scholar] [CrossRef]
- Hajek, P. Combining bag-of-words and sentiment features of annual reports to predict abnormal stock returns. Neural. Comput. Appl. 2018, 29, 343–358. [Google Scholar] [CrossRef] [Green Version]
- Dey, A.; Jenamani, M.; Thakkar, J.J. Lexical TF-IDF: An n-gram Feature Space for Cross-Domain Classification of Sentiment Reviews. In Proceedings of the Pattern Recognition and Machine Intelligence, Premi 2017, Kolkata, India, 5–8 December 2017; Volume 10597, pp. 380–386. [Google Scholar] [CrossRef]
- Kaur, S.; Sikka, G.; Awasthi, L.K. Sentiment Analysis Approach Based on N-gram and KNN Classifier. In Proceedings of the 2018 First International Conference on Secure Cyber Computing and Communications (Icsccc 2018), Jalandhar, India, 15–17 December 2018; pp. 13–16. [Google Scholar]
- Manek, A.S.; Shenoy, P.D.; Mohan, M.C.; Venugopal, K.R. Aspect term extraction for sentiment analysis in large movie reviews using Gini Index feature selection method and SVM classifier. World Wide Web (WWW) 2017, 20, 135–154. [Google Scholar] [CrossRef] [Green Version]
- Rout, J.K.; Choo, K.K.R.; Dash, A.K.; Bakshi, S.; Jena, S.K.; Williams, K.L. A model for sentiment and emotion analysis of unstructured social media text. Electron. Commer. Res. 2018, 18, 181–199. [Google Scholar] [CrossRef]
- Huq, M.R.; Ali, A.; Rahman, A. Sentiment Analysis on Twitter Data using KNN and SVM. Int. J. Adv. Comput. Sci. Appl. 2017, 8, 19–25. [Google Scholar]
- Zhang, Y.H.; Wang, Q.Q.; Li, Y.L.; Wu, X.D. Sentiment Classification Based on Piecewise Pooling Convolutional Neural Network. Comput. Mater. Contin. 2018, 56, 285–297. [Google Scholar] [CrossRef]
- Tang, H.L.; Mi, Y.; Xue, F.; Cao, Y. An Integration Model Based on Graph Convolutional Network for Text Classification. IEEE Access 2020, 8, 148865–148876. [Google Scholar] [CrossRef]
- Li, Q.; Li, P.F.; Mao, K.Z.; Lo, E.Y.M. Improving convolutional neural network for text classification by recursive data pruning. Neurocomputing 2020, 414, 143–152. [Google Scholar] [CrossRef]
- Usama, M.; Ahmad, B.; Song, E.; Hossain, M.S.; Alrashoud, M.; Muhammad, G. Attention-based sentiment analysis using convolutional and recurrent neural network. Future Gener. Comput. Syst. 2020, 113, 571–578. [Google Scholar] [CrossRef]
- Fu, X.H.; Yang, J.Y.; Li, J.Q.; Fang, M.; Wang, H.H. Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis. IEEE Access 2018, 6, 71884–71891. [Google Scholar] [CrossRef]
- Al-Smadi, M.; Talafha, B.; Al-Ayyoub, M.; Jararweh, Y. Using long short-term memory deep neural networks for aspect-based sentiment analysis of Arabic reviews. Int. J. Mach. Learn. Cybern. 2019, 10, 2163–2175. [Google Scholar] [CrossRef]
- Da’u, A.; Salim, N. Aspect extraction on user textual reviews using multi-channel convolutional neural network. Peerj Comput. Sci. 2019, 10, 2163–2175. [Google Scholar] [CrossRef] [Green Version]
- Gan, C.Q.; Wang, L.; Zhang, Z.F. Multi-entity sentiment analysis using self-attention based hierarchical dilated convolutional neural network. Future Gener. Comput. Syst. 2020, 112, 116–125. [Google Scholar] [CrossRef]
- Jelodar, H.; Wang, Y.L.; Orji, R.; Huang, S.C. Deep Sentiment Classification and Topic Discovery on Novel Coronavirus or COVID-19 Online Discussions: NLP Using LSTM Recurrent Neural Network Approach. IEEE J. Biomed. Health 2020, 24, 2733–2742. [Google Scholar] [CrossRef] [PubMed]
- Priyadarshini, I.; Cotton, C. A novel LSTM-CNN-grid search-based deep neural network for sentiment analysis. J. Supercomput. 2021, 77, 13911–13932. [Google Scholar] [CrossRef]
- Yin, R.C.; Li, P.; Wang, B. Sentiment Lexical-Augmented Convolutional Neural Networks for Sentiment Analysis. In Proceedings of the 2017 IEEE Second International Conference on Data Science in Cyberspace (Dsc), Shenzhen, China, 26–29 June 2017; pp. 630–635. [Google Scholar] [CrossRef]
- Wang, X.Y.; Li, F.; Zhang, Z.Q.; Xu, G.L.; Zhang, J.Y.; Sun, X. A unified position-aware convolutional neural network for aspect based sentiment analysis. Neurocomputing 2021, 450, 91–103. [Google Scholar] [CrossRef]
- Gan, C.Q.; Wang, L.; Zhang, Z.F.; Wang, Z.Y. Sparse attention based separable dilated convolutional neural network for targeted sentiment analysis. Knowl.-Based Syst. 2020, 188, 104827. [Google Scholar] [CrossRef]
- Zhao, P.L.; Hou, L.L.; Wu, O. Modeling sentiment dependencies with graph convolutional networks for aspect-level sentiment classification. Knowl.-Based Syst. 2020, 193, 105443. [Google Scholar] [CrossRef] [Green Version]
- Zhou, J.; Huang, J.X.; Hu, Q.V.; He, L. SK-GCN: Modeling Syntax and Knowledge via Graph Convolutional Network for aspect-level sentiment classification. Knowl.-Based Syst. 2020, 205, 106292. [Google Scholar] [CrossRef]
- Lu, Q.; Zhu, Z.F.; Zhang, G.Y.; Kang, S.Y.; Liu, P.Y. Aspect-gated graph convolutional networks for aspect-based sentiment analysis. Appl. Intell. 2021, 51, 4408–4419. [Google Scholar] [CrossRef]
- Ma, Y.K.; Peng, H.Y.; Cambria, E. Targeted Aspect-Based Sentiment Analysis via Embedding Commonsense Knowledge into an Attentive LSTM. In Proceedings of the Thirty-Second Aaai Conference on Artificial Intelligence/Thirtieth Innovative Applications of Artificial Intelligence Conference/Eighth Aaai Symposium on Educational Advances in Artificial Intelligence, New Orleans, LA, USA, 2–7 February 2018; pp. 5876–5883. [Google Scholar]
- Bao, L.X.; Lambert, P.; Badia, T. Attention and Lexicon Regularized LSTM for Aspect-based Sentiment Analysis. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (Acl 2019): Student Research Workshop, Florence, Italy, 28 July–2 August 2019; pp. 253–259. [Google Scholar]
- Ahmed, M.; Chen, Q.; Li, Z.H. Constructing domain-dependent sentiment dictionary for sentiment analysis. Neural Comput. Appl. 2020, 32, 14719–14732. [Google Scholar] [CrossRef]
- Hassan, J.; Shoaib, U. Multi-class Review Rating Classification using Deep Recurrent Neural Network. Neural Process. Lett. 2020, 51, 1031–1048. [Google Scholar] [CrossRef]
- Cambria, E.; Poria, S.; Hazarika, D.; Kwok, K. SenticNet 5: Discovering Conceptual Primitives for Sentiment Analysis by Means of Context Embeddings. In Proceedings of the Thirty-Second Aaai Conference on Artificial Intelligence/Thirtieth Innovative Applications of Artificial Intelligence Conference/Eighth Aaai Symposium on Educational Advances in Artificial Intelligence, New Orleans, LA, USA, 2–7 February 2018; pp. 1795–1802. [Google Scholar]
- Cambria, E.; Li, Y.; Xing, F.Z.; Poria, S.; Kwok, K. SenticNet 6: Ensemble application of symbolic and subsymbolic AI for sentiment analysis. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management, New York, NY, USA, 19–23 October 2020; pp. 105–114. [Google Scholar]
- Wei, J.Y.; Liao, J.; Yang, Z.F.; Wang, S.G.; Zhao, Q. BiLSTM with Multi-Polarity Orthogonal Attention for Implicit Sentiment Analysis. Neurocomputing 2020, 383, 165–173. [Google Scholar] [CrossRef]
- Li, W.; Zhu, L.Y.; Shi, Y.; Guo, K.; Cambria, E. User reviews: Sentiment analysis using lexicon integrated two-channel CNN-LSTM family models. Appl. Soft Comput. 2020, 94, 106435. [Google Scholar] [CrossRef]
- Behera, R.K.; Jena, M.; Rath, S.K.; Misra, S. Co-LSTM: Convolutional LSTM model for sentiment analysis in social big data. Inf. Process. Manag. 2021, 58, 102435. [Google Scholar] [CrossRef]
- Zhang, Y.Z.; Song, D.W.; Li, X.; Zhang, P.; Wang, P.P.; Rong, L.; Yu, G.L.; Wang, B. A Quantum -Like multimodal network framework for modeling interaction dynamics in multiparty conversational sentiment analysis. Inf. Fusion 2020, 62, 14–31. [Google Scholar] [CrossRef]
- Zhao, N.; Gao, H.; Wen, X.; Li, H. Combination of Convolutional Neural Network and Gated Recurrent Unit for Aspect-Based Sentiment Analysis. IEEE Access 2021, 9, 15561–15569. [Google Scholar] [CrossRef]
- Basiri, M.E.; Nemati, S.; Abdar, M.; Cambria, E.; Acharya, U.R. ABCDM: An Attention-based Bidirectional CNN-RNN Deep Model for sentiment analysis. Future Gener. Comput. Syst. 2021, 115, 279–294. [Google Scholar] [CrossRef]
- Gan, C.Q.; Feng, Q.D.; Zhang, Z.F. Scalable multi-channel dilated CNN-BiLSTM model with attention mechanism for Chinese textual sentiment analysis. Future Gener. Comput. Syst. 2021, 118, 297–309. [Google Scholar] [CrossRef]
- Feng, Y.; Cheng, Y. Short Text Sentiment Analysis Based on Multi-Channel CNN With Multi-Head Attention Mechanism. IEEE Access 2021, 9, 19854–19863. [Google Scholar] [CrossRef]
- Li, W.J.; Qi, F.; Tang, M.; Yu, Z.T. Bidirectional LSTM with self-attention mechanism and multi-channel features for sentiment classification. Neurocomputing 2020, 387, 63–77. [Google Scholar] [CrossRef]
- Zhang, M.S.; Zhang, Y.; Che, W.X.; Liu, T. Character-Level Chinese Dependency Parsing. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, Baltimore, MD, USA, 22–27 June 2014; Volume 1, pp. 1326–1336. [Google Scholar]
- Liu, Y.; Lu, J.H.; Yang, J.; Mao, F. Sentiment analysis for e-commerce product reviews by deep learning model of Bert-BiGRU-Softmax. Math. Biosci. Eng. 2020, 17, 7819–7837. [Google Scholar] [CrossRef] [PubMed]
- Liu, P.; Qiu, X.; Huang, X. Recurrent Neural Network for Text Classification with Multi-Task Learning. In Proceedings of the IJCAI, New York, NY, USA, 9–16 July 2016; pp. 2873–2879. [Google Scholar]
- Tai, K.S.; Socher, R.; Manning, C.D. Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, China, 26–31 July 2015; Volume 1, pp. 1556–1566. [Google Scholar]
- Cho, K.; van Merrienboer, B.; Gulcehre, C.; Bahdanau, D.; Bougares, F.; Schwenk, H.; Bengio, Y. Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation. arXiv 2014, arXiv:1406.1078. [Google Scholar]
- Chen, C.T.; Zhuo, R.; Ren, J.T. Gated recurrent neural network with sentimental relations for sentiment classification. Inf. Sci. 2019, 502, 268–278. [Google Scholar] [CrossRef]
- Zhang, X.; Zhao, J.B.; Yann, L.C. Character-level Convolutional Networks for Text Classification. In Proceedings of the Advances in Neural Information Processing Systems 28 (Nips 2015), Montreal, QC, Canada, 7–12 December 2015; Volume 28. [Google Scholar]
- Yang, L.; Li, Y.; Wang, J.; Sherratt, R.S. Sentiment Analysis for E-Commerce Product Reviews in Chinese Based on Sentiment Lexicon and Deep Learning. IEEE Access 2020, 8, 23522–23530. [Google Scholar] [CrossRef]
- Wang, B.; Shan, D.; Fan, A.; Liu, L.; Gao, J. A Sentiment Classification Method of Web Social Media Based on Multi-dimension and Multi-level Modeling. IEEE Trans. Ind. Inform. 2021, 18, 1240–1249. [Google Scholar] [CrossRef]
Sentiment Polarity | Number | Avg_len | |
---|---|---|---|
Train | Test | ||
positive | 36,450 | 12,150 | 32.58 |
negative | 36,450 | 12,150 |
Length | 10 | 20 | 30 | 40 | 50 | >50 |
Number | 30,910 | 22,973 | 13,820 | 8984 | 5492 | 17,821 |
Probability | 0.31 | 0.23 | 0.14 | 0.09 | 0.05 | 0.18 |
Methods | Accuracy | Precision | Recall | F1 |
---|---|---|---|---|
LSTM | 0.932 ± 0.001 | 0.931 ± 0.002 | 0.934 ± 0.003 | 0.933 ± 0.001 |
2-layer LSTM | 0.935 ± 0.001 | 0.932 ± 0.002 | 0.933 ± 0.002 | 0.932 ± 0.001 |
BiLSTM | 0.931 ± 0.000 | 0.932 ± 0.004 | 0.932 ± 0.005 | 0.932 ± 0.000 |
2-layer BiLSTM | 0.930 ± 0.001 | 0.932 ± 0.003 | 0.929 ± 0.003 | 0.931 ± 0.001 |
GRU | 0.867 ± 0.002 | 0.871 ± 0.003 | 0.864 ± 0.006 | 0.868 ± 0.002 |
BiGRU | 0.930 ± 0.001 | 0.928 ± 0.002 | 0.933 ± 0.003 | 0.931 ± 0.001 |
Character-level ConvNets | 0.928 ± 0.000 | 0.929 ± 0.004 | 0.930 ± 0.006 | 0.929 ± 0.001 |
SLCABG | 0.934 ± 0.000 | 0.931 ± 0.005 | 0.937 ± 0.006 | 0.934 ± 0.000 |
MDMLSM | 0.930 ± 0.001 | 0.931 ± 0.002 | 0.929 ± 0.003 | 0.930 ± 0.001 |
MCNN-IFGS (word-based) | 0.970 ± 0.004 | 0.978 ± 0.004 | 0.966 ± 0.006 | 0.972 ± 0.004 |
MCNN-IFGS (Ours) | 0.972 ± 0.003 | 0.974 ± 0.006 | 0.975 ± 0.008 | 0.974 ± 0.003 |
Learning Rate | Accuracy | Precision | Recall | F1 |
---|---|---|---|---|
0.001 | 0.972 ± 0.003 | 0.974 ± 0.006 | 0.975 ± 0.008 | 0.974 ± 0.003 |
0.002 | 0.965 ± 0.006 | 0.971 ± 0.007 | 0.966 ± 0.009 | 0.968 ± 0.005 |
0.003 | 0.969 ± 0.003 | 0.971 ± 0.002 | 0.973 ± 0.005 | 0.972 ± 0.002 |
0.004 | 0.955 ± 0.025 | 0.961 ± 0.012 | 0.957 ± 0.040 | 0.959 ± 0.024 |
0.005 | 0.933 ± 0.029 | 0.923 ± 0.031 | 0.957 ± 0.024 | 0.940 ± 0.026 |
0.006 | 0.913 ± 0.066 | 0.923 ± 0.055 | 0.917 ± 0.071 | 0.920 ± 0.062 |
0.007 | 0.890 ± 0.053 | 0.911 ± 0.047 | 0.886 ± 0.060 | 0.898 ± 0.050 |
0.008 | 0.818 ± 0.162 | 0.848 ± 0.121 | 0.777 ± 0.251 | 0.934 ± 0.209 |
Dropout | Accuracy | Precision | Recall | F1 |
---|---|---|---|---|
0.2 | 0.964 ± 0.002 | 0.962 ± 0.002 | 0.973 ± 0.004 | 0.968 ± 0.002 |
0.3 | 0.962 ± 0.002 | 0.967 ± 0.007 | 0.965 ± 0.011 | 0.966 ± 0.002 |
0.4 | 0.966 ± 0.002 | 0.963 ± 0.006 | 0.975 ± 0.005 | 0.969 ± 0.002 |
0.5 | 0.967 ± 0.004 | 0.971 ± 0.010 | 0.968 ± 0.007 | 0.970 ± 0.003 |
0.6 | 0.968 ± 0.005 | 0.971 ± 0.004 | 0.971 ± 0.013 | 0.971 ± 0.005 |
0.7 | 0.967 ± 0.002 | 0.969 ± 0.005 | 0.971 ± 0.007 | 0.970 ± 0.002 |
0.8 | 0.972 ± 0.003 | 0.974 ± 0.006 | 0.975 ± 0.008 | 0.974 ± 0.003 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Wang, L.; Meng, Z. Multichannel Two-Dimensional Convolutional Neural Network Based on Interactive Features and Group Strategy for Chinese Sentiment Analysis. Sensors 2022, 22, 714. https://doi.org/10.3390/s22030714
Wang L, Meng Z. Multichannel Two-Dimensional Convolutional Neural Network Based on Interactive Features and Group Strategy for Chinese Sentiment Analysis. Sensors. 2022; 22(3):714. https://doi.org/10.3390/s22030714
Chicago/Turabian StyleWang, Lin, and Zuqiang Meng. 2022. "Multichannel Two-Dimensional Convolutional Neural Network Based on Interactive Features and Group Strategy for Chinese Sentiment Analysis" Sensors 22, no. 3: 714. https://doi.org/10.3390/s22030714
APA StyleWang, L., & Meng, Z. (2022). Multichannel Two-Dimensional Convolutional Neural Network Based on Interactive Features and Group Strategy for Chinese Sentiment Analysis. Sensors, 22(3), 714. https://doi.org/10.3390/s22030714