A CNN-BiLSTM Model for Document-Level Sentiment Analysis
Document-level sentiment analysis is a challenging task given the large size of the text, which leads to an abundance of words and opinions, at times contradictory, in the same document. This analysis is particularly useful in analyzing press articles and blog posts about a particular product or com...
Saved in:
| Published in | Machine learning and knowledge extraction Vol. 1; no. 3; pp. 832 - 847 |
|---|---|
| Main Authors | , , , |
| Format | Journal Article |
| Language | English |
| Published |
01.09.2019
|
| Online Access | Get full text |
| ISSN | 2504-4990 2504-4990 |
| DOI | 10.3390/make1030048 |
Cover
| Abstract | Document-level sentiment analysis is a challenging task given the large size of the text, which leads to an abundance of words and opinions, at times contradictory, in the same document. This analysis is particularly useful in analyzing press articles and blog posts about a particular product or company, and it requires a high concentration, especially when the topic being discussed is sensitive. Nevertheless, most existing models and techniques are designed to process short text from social networks and collaborative platforms. In this paper, we propose a combination of Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) models, with Doc2vec embedding, suitable for opinion analysis in long texts. The CNN-BiLSTM model is compared with CNN, LSTM, BiLSTM and CNN-LSTM models with Word2vec/Doc2vec embeddings. The Doc2vec with CNN-BiLSTM model was applied on French newspapers articles and outperformed the other models with 90.66% accuracy. |
|---|---|
| AbstractList | Document-level sentiment analysis is a challenging task given the large size of the text, which leads to an abundance of words and opinions, at times contradictory, in the same document. This analysis is particularly useful in analyzing press articles and blog posts about a particular product or company, and it requires a high concentration, especially when the topic being discussed is sensitive. Nevertheless, most existing models and techniques are designed to process short text from social networks and collaborative platforms. In this paper, we propose a combination of Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) models, with Doc2vec embedding, suitable for opinion analysis in long texts. The CNN-BiLSTM model is compared with CNN, LSTM, BiLSTM and CNN-LSTM models with Word2vec/Doc2vec embeddings. The Doc2vec with CNN-BiLSTM model was applied on French newspapers articles and outperformed the other models with 90.66% accuracy. |
| Author | Rhanoui, Maryem Mikram, Mounia Yousfi, Siham Barzali, Soukaina |
| Author_xml | – sequence: 1 givenname: Maryem orcidid: 0000-0002-0147-8466 surname: Rhanoui fullname: Rhanoui, Maryem – sequence: 2 givenname: Mounia orcidid: 0000-0002-6647-5098 surname: Mikram fullname: Mikram, Mounia – sequence: 3 givenname: Siham orcidid: 0000-0002-0745-8327 surname: Yousfi fullname: Yousfi, Siham – sequence: 4 givenname: Soukaina surname: Barzali fullname: Barzali, Soukaina |
| BookMark | eNptj0tPwzAQhC1UJErpiT-QOzKsH3n4GMJTSsuh5RxtElsyJDGyA1L_PYngUCFOO7v6ZrRzThaDGzQhlwyuhVBw0-O7ZiAAZHZCljwGSaVSsDjSZ2QdwhsA8FRJBnJJsjwqtlt6a8vdfhNtXKu7yDgf3bnms9fDSEv9NZ12k7TzHuUDdodgwwU5NdgFvf6dK_L6cL8vnmj58vhc5CVtOMBIUw1cxkLFbWJaWXMwzfRNnWZxBlJJUSulUmY41lpJPtOACBizGhPGEyNW5Oont_EuBK9N9eFtj_5QMajm3tVR74lmf-jGjjhaN4webfev5xvsHlq- |
| CitedBy_id | crossref_primary_10_3390_make6010023 crossref_primary_10_2196_22916 crossref_primary_10_3390_su13169087 crossref_primary_10_3390_app15052345 crossref_primary_10_1109_ACCESS_2022_3210182 crossref_primary_10_1007_s10462_023_10651_9 crossref_primary_10_1016_j_eswa_2022_117575 crossref_primary_10_1016_j_compeleceng_2022_107978 crossref_primary_10_1016_j_eswa_2021_115265 crossref_primary_10_1007_s10666_023_09920_2 crossref_primary_10_1016_j_neucom_2024_127732 crossref_primary_10_4018_IJACI_331744 crossref_primary_10_3390_informatics10040086 crossref_primary_10_1016_j_ijhcs_2021_102761 crossref_primary_10_3233_WEB_230063 crossref_primary_10_1007_s10207_024_00945_6 crossref_primary_10_1109_ACCESS_2021_3073728 crossref_primary_10_1002_spe_2853 crossref_primary_10_1145_3586075 crossref_primary_10_1371_journal_pone_0317355 crossref_primary_10_1016_j_seta_2023_103263 crossref_primary_10_1109_ACCESS_2024_3394930 crossref_primary_10_1016_j_copbio_2023_102968 crossref_primary_10_1109_ACCESS_2021_3067844 crossref_primary_10_1109_ACCESS_2021_3096739 crossref_primary_10_1109_ACCESS_2025_3541494 crossref_primary_10_3390_w17010059 crossref_primary_10_1007_s10489_022_04060_8 crossref_primary_10_1109_ACCESS_2022_3152828 crossref_primary_10_1109_ACCESS_2021_3072900 crossref_primary_10_7717_peerj_cs_2592 crossref_primary_10_1051_shsconf_202316304007 crossref_primary_10_3390_app9194156 crossref_primary_10_1109_ACCESS_2024_3479774 crossref_primary_10_1142_S0218126623502924 crossref_primary_10_1007_s40614_023_00394_x crossref_primary_10_3390_sym13030406 crossref_primary_10_1016_j_eswa_2023_119611 crossref_primary_10_1016_j_ijhm_2023_103647 crossref_primary_10_3390_mi14112050 crossref_primary_10_7717_peerj_cs_1414 crossref_primary_10_7717_peerj_cs_2349 crossref_primary_10_3390_computers14020036 crossref_primary_10_3390_s24041162 crossref_primary_10_3390_electronics14030434 crossref_primary_10_3390_sym13081517 crossref_primary_10_29109_gujsc_1111884 crossref_primary_10_1016_j_procs_2024_04_218 crossref_primary_10_1093_comjnl_bxab118 crossref_primary_10_1155_2022_4914792 crossref_primary_10_1007_s00521_022_07248_8 crossref_primary_10_3390_machines10050336 crossref_primary_10_4018_IJITSA_324100 crossref_primary_10_1007_s00500_020_05290_z crossref_primary_10_3390_app11188613 crossref_primary_10_1016_j_atech_2024_100573 crossref_primary_10_1007_s10462_022_10215_3 crossref_primary_10_3233_JIFS_237653 crossref_primary_10_1049_cth2_12463 crossref_primary_10_1007_s00034_025_02999_w crossref_primary_10_51583_IJLTEMAS_2024_130721 crossref_primary_10_3390_math9212722 crossref_primary_10_1007_s13735_023_00308_2 crossref_primary_10_31083_j_fbl2812322 crossref_primary_10_3390_electronics11121906 crossref_primary_10_1007_s12559_021_09836_7 crossref_primary_10_61186_jsdp_20_2_3 crossref_primary_10_3390_electronics12061298 crossref_primary_10_1007_s42979_023_01913_y crossref_primary_10_1016_j_inffus_2025_103000 crossref_primary_10_1016_j_engappai_2024_108231 crossref_primary_10_3390_math11102335 crossref_primary_10_3390_app13074550 crossref_primary_10_1186_s40537_025_01064_2 crossref_primary_10_1007_s10614_024_10566_9 crossref_primary_10_1007_s10462_024_10795_2 crossref_primary_10_3390_make2030016 crossref_primary_10_1016_j_energy_2024_133596 crossref_primary_10_1155_2021_5541134 crossref_primary_10_1016_j_iswa_2022_200090 crossref_primary_10_1016_j_rineng_2023_101109 crossref_primary_10_1016_j_engappai_2023_107305 crossref_primary_10_1016_j_jisa_2024_103953 crossref_primary_10_1038_s41598_023_41485_8 crossref_primary_10_32604_csse_2023_029953 crossref_primary_10_1007_s41870_024_01882_2 crossref_primary_10_1002_cpe_7538 crossref_primary_10_1007_s00500_023_08639_2 crossref_primary_10_3390_app13063915 crossref_primary_10_3390_make6020041 crossref_primary_10_1108_IJICC_04_2024_0184 crossref_primary_10_1007_s13246_022_01189_1 crossref_primary_10_1109_ACCESS_2024_3487752 crossref_primary_10_1109_TICPS_2024_3477433 crossref_primary_10_1007_s10115_025_02365_x crossref_primary_10_1007_s44196_024_00589_3 crossref_primary_10_1007_s11042_022_13810_2 crossref_primary_10_1088_2632_2153_abd614 crossref_primary_10_1007_s42979_021_00977_y crossref_primary_10_1016_j_jksuci_2020_10_017 crossref_primary_10_1007_s11219_025_09714_7 crossref_primary_10_1007_s10115_024_02256_7 crossref_primary_10_3390_systems11030148 crossref_primary_10_1007_s41870_023_01395_4 crossref_primary_10_1007_s11042_024_18316_7 crossref_primary_10_1016_j_eswa_2022_117581 crossref_primary_10_1080_17449057_2022_2031509 crossref_primary_10_3390_electronics13010004 crossref_primary_10_1016_j_cja_2022_05_005 crossref_primary_10_1016_j_knosys_2021_107073 crossref_primary_10_1007_s41019_021_00165_1 crossref_primary_10_21923_jesd_1350375 crossref_primary_10_1109_ACCESS_2021_3110143 crossref_primary_10_1109_ACCESS_2023_3276757 crossref_primary_10_1016_j_dajour_2022_100111 crossref_primary_10_3390_buildings15050763 crossref_primary_10_1007_s11042_024_18124_z crossref_primary_10_24018_ejece_2023_7_6_583 crossref_primary_10_1109_ACCESS_2023_3259107 crossref_primary_10_2139_ssrn_4165241 crossref_primary_10_1109_ACCESS_2021_3121791 crossref_primary_10_1016_j_engappai_2023_106041 crossref_primary_10_1007_s10462_023_10555_8 |
| Cites_doi | 10.1007/s13278-012-0057-9 10.1007/978-3-642-35236-2_62 10.18653/v1/W16-1609 10.1007/978-3-319-68121-4_17 10.18653/v1/D16-1058 10.1007/978-3-7908-2604-3_16 10.3115/v1/D14-1162 10.3115/v1/S14-2033 10.1007/978-3-642-35289-8_3 10.1162/neco.1997.9.8.1735 10.21437/Interspeech.2012-65 10.1002/widm.1253 10.3115/v1/P14-2009 10.1016/j.ipm.2016.12.002 10.1016/j.eswa.2018.06.052 10.18653/v1/N16-1174 10.1145/2766462.2767830 10.1145/1651461.1651468 10.1145/945645.945658 10.1016/j.ipm.2013.08.006 10.1109/CTS.2012.6261103 10.1609/aaai.v29i1.9513 10.1109/78.650093 10.1016/j.giq.2012.06.005 10.1109/ACCESS.2017.2776930 10.3115/v1/D14-1181 10.14313/JAMRIS/3-2020/34 10.1162/tacl_a_00051 10.1016/j.neucom.2018.04.045 10.18653/v1/W18-5408 10.3390/make1010014 10.3115/v1/P15-1150 10.1109/UEMCON.2017.8249013 10.1561/1500000011 |
| ContentType | Journal Article |
| DBID | AAYXX CITATION |
| DOI | 10.3390/make1030048 |
| DatabaseName | CrossRef |
| DatabaseTitle | CrossRef |
| DatabaseTitleList | CrossRef |
| DeliveryMethod | fulltext_linktorsrc |
| EISSN | 2504-4990 |
| EndPage | 847 |
| ExternalDocumentID | 10_3390_make1030048 |
| GroupedDBID | AADQD AAFWJ AAYXX AFPKN AFZYC ALMA_UNASSIGNED_HOLDINGS CITATION GROUPED_DOAJ MODMG M~E OK1 |
| ID | FETCH-LOGICAL-c200t-7e0245395d6fd4b20fc250b785804943b99971f2abe9427e020aa0a51ba6126f3 |
| ISSN | 2504-4990 |
| IngestDate | Thu Apr 24 23:10:13 EDT 2025 Tue Jul 01 03:11:06 EDT 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 3 |
| Language | English |
| License | https://creativecommons.org/licenses/by/4.0 |
| LinkModel | OpenURL |
| MergedId | FETCHMERGED-LOGICAL-c200t-7e0245395d6fd4b20fc250b785804943b99971f2abe9427e020aa0a51ba6126f3 |
| ORCID | 0000-0002-6647-5098 0000-0002-0147-8466 0000-0002-0745-8327 |
| OpenAccessLink | https://www.mdpi.com/2504-4990/1/3/48/pdf?version=1564047828 |
| PageCount | 16 |
| ParticipantIDs | crossref_primary_10_3390_make1030048 crossref_citationtrail_10_3390_make1030048 |
| ProviderPackageCode | CITATION AAYXX |
| PublicationCentury | 2000 |
| PublicationDate | 2019-09-01 |
| PublicationDateYYYYMMDD | 2019-09-01 |
| PublicationDate_xml | – month: 09 year: 2019 text: 2019-09-01 day: 01 |
| PublicationDecade | 2010 |
| PublicationTitle | Machine learning and knowledge extraction |
| PublicationYear | 2019 |
| References | ref_13 ref_12 ref_11 ref_19 ref_18 ref_17 ref_15 Sobkowicz (ref_31) 2012; 29 Missen (ref_36) 2013; 3 Gamal (ref_16) 2019; 1 Rao (ref_38) 2018; 308 ref_25 ref_24 ref_21 ref_20 ref_29 ref_28 ref_27 ref_26 Pang (ref_7) 2008; 2 Fu (ref_39) 2018; 113 Jianqiang (ref_22) 2018; 6 Uysal (ref_23) 2014; 50 ref_35 ref_33 ref_32 ref_30 ref_37 Schuster (ref_14) 1997; 45 Hu (ref_34) 2017; 53 Hochreiter (ref_10) 1997; 9 Bojanowski (ref_50) 2017; 5 ref_47 ref_46 ref_45 ref_44 ref_43 ref_42 ref_41 Goldberg (ref_6) 2017; 10 ref_40 ref_3 ref_2 Liu (ref_1) 2012; 5 ref_49 ref_48 ref_9 ref_8 ref_5 ref_4 |
| References_xml | – ident: ref_9 – volume: 3 start-page: 107 year: 2013 ident: ref_36 article-title: Opinion mining: reviewed from word to document level publication-title: Soc. Netw. Anal. Min. doi: 10.1007/s13278-012-0057-9 – ident: ref_24 doi: 10.1007/978-3-642-35236-2_62 – ident: ref_40 doi: 10.18653/v1/W16-1609 – ident: ref_5 – ident: ref_26 – ident: ref_29 doi: 10.1007/978-3-319-68121-4_17 – ident: ref_47 doi: 10.18653/v1/D16-1058 – ident: ref_44 doi: 10.1007/978-3-7908-2604-3_16 – ident: ref_42 – ident: ref_49 doi: 10.3115/v1/D14-1162 – ident: ref_20 doi: 10.3115/v1/S14-2033 – ident: ref_8 – ident: ref_27 – ident: ref_43 doi: 10.1007/978-3-642-35289-8_3 – volume: 5 start-page: 1 year: 2012 ident: ref_1 article-title: Sentiment analysis and opinion mining publication-title: Synth. Lect. Hum. Lang. Technol. – volume: 9 start-page: 1735 year: 1997 ident: ref_10 article-title: Long short-term memory publication-title: Neural Comput. doi: 10.1162/neco.1997.9.8.1735 – ident: ref_4 doi: 10.21437/Interspeech.2012-65 – ident: ref_45 – ident: ref_17 doi: 10.1002/widm.1253 – ident: ref_19 doi: 10.3115/v1/P14-2009 – volume: 53 start-page: 436 year: 2017 ident: ref_34 article-title: Opinion mining from online hotel reviews–A text summarization approach publication-title: Inf. Process. Manag. doi: 10.1016/j.ipm.2016.12.002 – volume: 113 start-page: 33 year: 2018 ident: ref_39 article-title: Bag of meta-words: A novel method to represent document for the sentiment classification publication-title: Expert Syst. Appl. doi: 10.1016/j.eswa.2018.06.052 – ident: ref_48 doi: 10.18653/v1/N16-1174 – ident: ref_21 doi: 10.1145/2766462.2767830 – ident: ref_30 – ident: ref_32 doi: 10.1145/1651461.1651468 – ident: ref_2 doi: 10.1145/945645.945658 – volume: 50 start-page: 104 year: 2014 ident: ref_23 article-title: The impact of preprocessing on text classification publication-title: Inf. Process. Manag. doi: 10.1016/j.ipm.2013.08.006 – ident: ref_11 – ident: ref_46 doi: 10.1109/CTS.2012.6261103 – ident: ref_13 doi: 10.1609/aaai.v29i1.9513 – volume: 10 start-page: 1 year: 2017 ident: ref_6 article-title: Neural network methods for natural language processing publication-title: Synth. Lect. Hum. Lang. Technol. – volume: 45 start-page: 2673 year: 1997 ident: ref_14 article-title: Bidirectional recurrent neural networks publication-title: IEEE Trans. Signal Process. doi: 10.1109/78.650093 – ident: ref_37 – volume: 29 start-page: 470 year: 2012 ident: ref_31 article-title: Opinion mining in social media: Modeling, simulating, and forecasting political opinions in the web publication-title: Gov. Inf. Q. doi: 10.1016/j.giq.2012.06.005 – ident: ref_18 – volume: 6 start-page: 23253 year: 2018 ident: ref_22 article-title: Deep convolution neural networks for Twitter sentiment analysis publication-title: IEEE Access doi: 10.1109/ACCESS.2017.2776930 – ident: ref_25 – ident: ref_33 – ident: ref_12 – ident: ref_3 doi: 10.3115/v1/D14-1181 – ident: ref_35 doi: 10.14313/JAMRIS/3-2020/34 – volume: 5 start-page: 135 year: 2017 ident: ref_50 article-title: Enriching word vectors with subword information publication-title: Trans. Assoc. Comput. Linguist. doi: 10.1162/tacl_a_00051 – volume: 308 start-page: 49 year: 2018 ident: ref_38 article-title: LSTM with sentence representations for document-level sentiment classification publication-title: Neurocomputing doi: 10.1016/j.neucom.2018.04.045 – ident: ref_41 doi: 10.18653/v1/W18-5408 – volume: 1 start-page: 224 year: 2019 ident: ref_16 article-title: Analysis of Machine Learning Algorithms for Opinion Mining in Different Domains publication-title: Mach. Learn. Knowl. Extr. doi: 10.3390/make1010014 – ident: ref_15 doi: 10.3115/v1/P15-1150 – ident: ref_28 doi: 10.1109/UEMCON.2017.8249013 – volume: 2 start-page: 1 year: 2008 ident: ref_7 article-title: Opinion mining and sentiment analysis publication-title: Found. Trends Inf. Retr. doi: 10.1561/1500000011 |
| SSID | ssj0002794104 |
| Score | 2.4814684 |
| Snippet | Document-level sentiment analysis is a challenging task given the large size of the text, which leads to an abundance of words and opinions, at times... |
| SourceID | crossref |
| SourceType | Enrichment Source Index Database |
| StartPage | 832 |
| Title | A CNN-BiLSTM Model for Document-Level Sentiment Analysis |
| Volume | 1 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV07T8MwELagLCwIBIi3MnQCGRqnaZKxrUAIkS4tEltlO45aFVJUkoEO_HbOjzih6lBYoso6R0nvy_l8vu8OoWbAw45PuYvTIEoxrHgJjhgNsd9JPS44gUVNcofjQefxpf306r9WxwWKXZKzW75cyyv5j1ZhDPQqWbJ_0Ky9KQzAb9AvXEHDcN1Ix92b_mCAe9Pn4ShWXc0UGRGcYl7IoB9-lhlBYA6yXNXwtxVI6h5prLIpRdk-QjMWbaTtBmz3QnMf7NHMhGbzYmqIPl8m9KtohbOFhlcMJmRKawblM1Xyw-mEWvEeXSypJmcP58WMmj7eZQjCrXKsjKWSZdAwbJ30AYtYM1aa2hqivJrZDE2M06zAugbnqnH3vEhmQ77TmZC90Vq6QufvEtorS5tNOIStjpw-rk3eRjskCNTRfvxdxeUIGChXdZ20b6BpnXL-XW1-zZGpeSSjfbRnthJOV-PiAG2J7BCFXafChKMw4QAmnN-YcCwmnBITR-jl4X7Uf8SmOwbm8A3kOBDy1NyL_KSTJm1GWimHJ2ZB6Iey5o_HwPUP3JRQJqI2kdItSlvUdxkFrxY-xWPUyOaZOEFOFMC-FTy1hHBp09PQE-AaCiLDAcwnySm6Ll91zE3peNnB5G285n89RU0r_KErpqwTO9tM7BztVoi7QI18UYhLcANzdqXCJ1dKez_hK1vb |
| linkProvider | ISSN International Centre |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=A+CNN-BiLSTM+Model+for+Document-Level+Sentiment+Analysis&rft.jtitle=Machine+learning+and+knowledge+extraction&rft.au=Rhanoui%2C+Maryem&rft.au=Mikram%2C+Mounia&rft.au=Yousfi%2C+Siham&rft.au=Barzali%2C+Soukaina&rft.date=2019-09-01&rft.issn=2504-4990&rft.eissn=2504-4990&rft.volume=1&rft.issue=3&rft.spage=832&rft.epage=847&rft_id=info:doi/10.3390%2Fmake1030048&rft.externalDBID=n%2Fa&rft.externalDocID=10_3390_make1030048 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2504-4990&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2504-4990&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2504-4990&client=summon |