Chinese authorship identification based on fine-grained word feature

Abstract

Abstract: Most existing authorship identification models are primarily designed for English texts. However, due to the differences between Chinese and English in grammar and language elements, the English authorship identification models have large deviations when applied to Chinese text. To solve the problem of Chinese author identification, a model adapted to Chinese features is proposed, termed the Chinese author recognition model with fine-grained word features. The model uses parallel convolution to extract fine-grained features from 1 to 4 characters and combines with an attention mechanism for weight assignment. Finally, Chinese authorship identification is obtained by the classifier. The experimental results show that the accuracy of this model is average improved by 2.09%, 7.2%, and 6.71% on three Chinese author identification datasets compared with the baseline models of BERT, TextCNN, and RNN, separately. Therefore, this model has a high value in reality.

Key words: Chinese authorship identification, BERT, attention mechanism, parallel convolutional layers, fine-grained feature

CLC Number:

TP391.1

ZHAO Hong, ZHANG Chen-peng, WANG Ao-long, ZHANGYang. Chinese authorship identification based on fine-grained word feature[J]. Journal of Lanzhou University of Technology, 2025, 51(3): 99-106.

References

[1] FIFIELD D,FOLLAN T,LUNDE E.Unsupervised authorship attribution[EB/OL].[2015-03-26].https://arxiv.org/pdf/1503.07613.pdf.
[2] SEROUSSI Y,ZUKERMAN I,BOHNERT F.Authorship attribution with latent Dirichlet allocation[C]//Proceedings of the 15th Conference on Computational Natural Language Learning.Stroudsburg:ACL,2011:181-189.
[3] NAKOV P.Latent semantic analysis for German literature investigation[C]//International Conference on Computational Intelligence.Berlin:Springer,2001:834-841.
[4] 张洋,江铭虎.作者识别研究综述[J].自动化学报,2021,47(11):2501-2520.
[5] ARGAMON S.Scalability issues in authorship attribution[M].Antwerp:UPA University Press,2010:13-18.
[6] ANWAR W,BAJWA I S,RAMZAN S.Design and implementation of a machine learning-based authorship identification model[J].Scientific Programming,2019(3):9431073.
[7] OLIVEIRA L S,PAVELEC D,JUSTINO E.Author identification using stylometric features[J].Inteligencia Artificial,2007,11(36):59-65.
[8] SCHWARTZ R,TSUR O,RAPPOPORT A,et al.Authorship attribution of micro-messages[C]//Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing.Stroudsburg:ACL,2013:1880-1891.
[9] MARTIN-DEL-CAMPO-RODRíGUEZ C,ALVAREZ D A P,SIFUENTES C E M,et al.Authorship attribution through punctuation N-grams and averaged combination of SVM[C]//Proceedings of the CLEF.Lugano,Switzerland:[s.n.],2019:9-12.
[10] DIEDERICH J,KINDERMANN J,LEOPOLD E,et al.Authorship attribution with support vector machines[J].Applied Intelligence,2003,19(1):109-123.
[11] ZAREMBA W,SUTSKEVER I,VINYALS O.Recurrent neural network regularization[EB/OL].[2015-02-19].https://arxiv.org/abs/1409.2329.
[12] HOCHREITER S,SCHMIDHUBER J.Long short-term memory[J].Neural Computation,1997,9(8):1735-1780.
[13] CHUNG J,GULCEHRE C,CHO K H,et al.Empirical evaluation of gated recurrent neural networks on sequence modeling[EB/OL].[2014-12-11].https://arxiv.org/abs/1412.3555.
[14] BAGNALL D.Author identification using multi-headed recurrent neural networks[EB/OL].[2015-06-16].https://arxiv.org/ftp/arxiv/papers/1506/1506.04891.pdf.
[15] Al AZHAR I,AHMED S,ISLAM M S,et al.Identifying author in Bengali literature by Bi-LSTM with attention mechanism[C]//2021 24th International Conference on Computer and Information Technology (ICCIT).Dhaka, Bangladesh:IEEE,2021:1-6.
[16] QIAN C,HE T,ZHANG R.Deep learning based authorship identification[R].California:Stanford University,2017:1-9.
[17] RUDER S,GHAFFARI P,BRESLIN J G.Character-level and multi-channel convolutional neural networks for large-scale authorship attribution[J/OL].[2016-9-21].https://arxiv.org/abs/1609.06686.
[18] SCHERER D,MULLER A,BEHNKE S.Evaluation of pooling operations in convolutional architectures for object recognition[C]//International Conference on Artificial Neural Networks.Berlin:Springer,2010:92-101.
[19] ZHANG Y,WALLACE B.A sensitivity analysis of (and practitioners guide to) convolutional neural networks for sentence classification[EB/OL].[2015-10-13].https://arxiv.org/pdf/1510.03820.pdf.
[20] MNIH V,HEESS N,GRAVES A.Recurrent models of visual attention[EB/OL].[2014-08-24].https://arxiv.org/abs/1406.6247.
[21] 张洋,江铭虎.基于注意力机制的文本作者识别[J].计算机应用,2021,41(7):1897-1901.
[22] VAWANI A,SHAZEER N,PARMAR N,et al.Attention is all you need[C]//Advances in Neural Information Processing Systems.Cambridge,MA:MIT Press,2017:5998-6008.
[23] DEVLIN J,CHANG M W,LEE K,et al.BERT:Pre-training of deep bidirectional transformers for language understanding[EB/OL].[2018-10-11].https://https://arxiv.org/pdf/1810.04805.pdf.
[24] FABIEN M,VILLATORO-TELLO E,MOTICEK P,et al.BertAA:BERT fine-tuning for authorship attribution[C]//Proceedings of the 17th International Conference on Natural Language Processing (ICON).Stroudsburg:ACL,2020:127-137.
[25] BARLAS G,STAMATATOS E.Cross-domain authorship attribution using pre-trained language models[C]//IFIP International Conference on Artificial Intelligence Applications and Innovations.Berlin:Springer,2020:255-266.
[26] 李孟林,罗文华,李绍鸣.基于神经网络中文短文本作者识别研究[J].中国人民公安大学学报(自然科学版),2020,26(2):61-67.
[27] 祁瑞华,杨德礼,郭旭,等.基于多层面文体特征的博客作者身份识别研究[J].情报学报,2015,34(6):628-634.
[28] 张翼翔,芦天亮,李默.基于BERT-BiGRU-ATT的社交媒体用户身份识别研究[J].中国人民公安大学学报(自然科学版),2021,27(1):70-75.
[29] SCHUSTER M,PALIWAL K K.Bidirectional recurrent neural networks[J].IEEE Transactions on Signal Processing,1997,45(11):2673-2681.
[30] MIKOLOV T,CHEN K,CORRADO G,et al.Efficient estimation of word representations in vector space [EB/OL].[2013-01-06].https://arxiv.org/pdf/1301.3781.pdf.