Hybrid deep learning framework for robust time-series classification: Integrating inception modules with residual networks
Accurate time-series classification (TSC) remains a fundamental challenge in deep learning due to the complexity and variability of temporal patterns. While recurrent neural networks (RNNs) such as LSTM and GRU have shown promise in modeling sequential dependencies, they often suffer from limitation...
Uloženo v:
| Vydáno v: | Journal of algorithms & computational technology Ročník 19 |
|---|---|
| Hlavní autoři: | , , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
SAGE Publishing
01.06.2025
|
| ISSN: | 1748-3018, 1748-3026 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | Accurate time-series classification (TSC) remains a fundamental challenge in deep learning due to the complexity and variability of temporal patterns. While recurrent neural networks (RNNs) such as LSTM and GRU have shown promise in modeling sequential dependencies, they often suffer from limitations like vanishing gradients and high computational cost when handling long sequences. To overcome these issues, convolutional neural networks (CNNs), particularly the Inception architecture, have emerged as powerful alternatives due to their ability to capture multiscale local patterns efficiently. In this study, we propose InceptionResNet, a hybrid deep learning framework that integrates the residual learning mechanism of ResNet into the InceptionTime architecture. By replacing the fully convolutional network (FCN) shortcut module in InceptionFCN with ResNet-50, the model gains deeper representational capacity and improved gradient flow during training. We conduct extensive experiments on the UCR-85 benchmark dataset, comparing our model against state-of-the-art approaches, including InceptionTime, InceptionFCN, ResNet, FCN, and MLP. The results show that InceptionResNet achieves superior accuracy on 49 of 85 datasets, demonstrating its robustness and effectiveness in handling diverse and complex time series data. This work highlights the potential of integrating multiscale feature extraction and deep residual learning to advance the performance of TSC models in practical applications. |
|---|---|
| AbstractList | Accurate time-series classification (TSC) remains a fundamental challenge in deep learning due to the complexity and variability of temporal patterns. While recurrent neural networks (RNNs) such as LSTM and GRU have shown promise in modeling sequential dependencies, they often suffer from limitations like vanishing gradients and high computational cost when handling long sequences. To overcome these issues, convolutional neural networks (CNNs), particularly the Inception architecture, have emerged as powerful alternatives due to their ability to capture multiscale local patterns efficiently. In this study, we propose InceptionResNet, a hybrid deep learning framework that integrates the residual learning mechanism of ResNet into the InceptionTime architecture. By replacing the fully convolutional network (FCN) shortcut module in InceptionFCN with ResNet-50, the model gains deeper representational capacity and improved gradient flow during training. We conduct extensive experiments on the UCR-85 benchmark dataset, comparing our model against state-of-the-art approaches, including InceptionTime, InceptionFCN, ResNet, FCN, and MLP. The results show that InceptionResNet achieves superior accuracy on 49 of 85 datasets, demonstrating its robustness and effectiveness in handling diverse and complex time series data. This work highlights the potential of integrating multiscale feature extraction and deep residual learning to advance the performance of TSC models in practical applications. |
| Author | Kim Chi, Duong Thi Ngoc Thao, Nguyen Nguyen, Thanh Q. Minh Son, Tran Ba Mai Trang, Nguyen Thi |
| Author_xml | – sequence: 1 givenname: Duong Thi orcidid: 0000-0003-1744-3249 surname: Kim Chi fullname: Kim Chi, Duong Thi organization: Faculty of Engineering and Technology, Thu Dau Mot University, Binh Duong Province, Vietnam – sequence: 2 givenname: Nguyen Thi surname: Mai Trang fullname: Mai Trang, Nguyen Thi organization: Ho Chi Minh City Open University, Ho Chi Minh City, Vietnam – sequence: 3 givenname: Tran Ba surname: Minh Son fullname: Minh Son, Tran Ba organization: Institute of Information Technology and Digital Transformation, Thu Dau Mot University, Binh Duong Province, Vietnam – sequence: 4 givenname: Nguyen surname: Ngoc Thao fullname: Ngoc Thao, Nguyen organization: Dong Nai Provincial Police, Tan Tien Ward, Bien Hoa, Dong Nai, Vietnam – sequence: 5 givenname: Thanh Q. orcidid: 0000-0003-4898-091X surname: Nguyen fullname: Nguyen, Thanh Q. organization: Institute of Interdisciplinary Sciences, Nguyen Tat Thanh University, Ho Chi Minh City, Vietnam |
| BookMark | eNplkcFOwzAMhiM0JMbYA3DLCxTqpmlSbmgCNmkSFzhXbuKOjK6Zkk7TeHpahrjgi_3_sr-D_2s26XxHjN1Cegeg1D2oXIs0KzIJItdawgWbjl4ympO_GfQVm8e4TYcSmdIgpuxreaqDs9wS7XlLGDrXbXgTcEdHHz554wMPvj7EnvduR0mk4Chy02KMrnEGe-e7B77qetqEQQzHrjO0H22-8_bQDttH13_wQNHZA7a8o35Exxt22WAbaf7bZ-z9-eltsUzWry-rxeM6MSLVfUJ1mdu0RigRpVRlUyhVyNJIKyQWZJSAQiqT60FZSSYzCqAsG4QMlam1mLHVmWs9bqt9cDsMp8qjq34MHzYVht6ZlqoUhcDhNbYUJsdGaSuNIrCiJkkg1cCCM8sEH2Og5o8HaTVmUf3LQnwD0i6Ajg |
| Cites_doi | 10.1016/j.jrmge.2023.06.015 10.1186/s13677-023-00560-1 10.3390/math11030590 10.3389/fdata.2023.1282541 10.1016/j.ins.2023.119147 10.1002/oca.3122 10.1016/j.jeconom.2022.12.008 10.1109/ACCESS.2018.2814605 10.1007/s10489-024-05649-x 10.1007/s12145-024-01414-3 10.1007/s12541-024-01069-6 10.1080/17499518.2022.2138918 10.1007/s12517-021-08259-w 10.1007/s11269-022-03419-3 10.1007/s11063-022-10929-z 10.1063/5.0172297 10.1007/s10618-020-00710-y 10.7717/peerj-cs.1795 10.3390/s22010157 10.1007/s41066-023-00444-4 10.1201/9781003254515-8 10.1080/00273171.2023.2214787 10.1145/3649448 10.1007/s42979-020-00180-5 10.3390/en17091998 10.1007/s00521-024-09962-x 10.1016/j.jeconom.2023.105544 10.1016/j.compag.2023.107705 10.3390/a17020076 10.1109/ACCESS.2024.3369891 10.1609/aaai.v31i1.11231 10.1016/j.neucom.2021.02.046 10.1089/big.2022.0155 10.1109/ACCESS.2021.3091162 10.1016/j.procs.2024.03.007 10.1016/j.asoc.2022.109945 10.1109/ACCESS.2020.3023273 10.1016/j.ins.2023.119951 10.1007/s11042-021-11885-x 10.1007/s10618-021-00745-9 10.1016/j.eswa.2023.121177 |
| ContentType | Journal Article |
| DBID | AAYXX CITATION DOA |
| DOI | 10.1177/17483026251348851 |
| DatabaseName | CrossRef DOAJ Directory of Open Access Journals |
| DatabaseTitle | CrossRef |
| DatabaseTitleList | CrossRef |
| Database_xml | – sequence: 1 dbid: DOA name: Directory of Open Access Journals (DOAJ) url: https://www.doaj.org/ sourceTypes: Open Website |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Mathematics |
| EISSN | 1748-3026 |
| ExternalDocumentID | oai_doaj_org_article_0a33a327d93c4af78d5c7e1d3be5e157 10_1177_17483026251348851 |
| GroupedDBID | .4S .DC 0R~ 29J 4.4 54M 5GY 5VS 8G5 AAJPV AAOTM AATZT AAYXX ABAWP ABQXT ABUWG ACDXX ACGFS ACHEB ACROE ADBBV ADEBD ADMLS ADOGD AEDFJ AEWDL AFCOW AFFHD AFKRA AFKRG AFRWT AJUZI ALMA_UNASSIGNED_HOLDINGS AMVHM ARCSS AUTPY AYAKG AZQEC BCNDV BDDNI BENPR BPHCQ CCPQU CITATION CKLRP CS3 DWQXO EBS EDO EJD F5P GNUQQ GROUPED_DOAJ GUQSH H13 IL9 IPNFZ J8X J9A K.F KQ8 M2O MET MK~ MV1 O9- OK1 P2P PHGZM PHGZT PIMPY PQQKQ RIG ROL SAUOL SCDPB SCNPE SFC AASGM |
| ID | FETCH-LOGICAL-c308t-eb94d0ba19aa5579f677659c5d35a6ec731657c485a6d5ec2c71199fa12a7cb83 |
| IEDL.DBID | DOA |
| ISICitedReferencesCount | 0 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001505375500001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 1748-3018 |
| IngestDate | Fri Oct 03 12:42:47 EDT 2025 Sat Nov 29 07:47:54 EST 2025 |
| IsDoiOpenAccess | true |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c308t-eb94d0ba19aa5579f677659c5d35a6ec731657c485a6d5ec2c71199fa12a7cb83 |
| ORCID | 0000-0003-1744-3249 0000-0003-4898-091X |
| OpenAccessLink | https://doaj.org/article/0a33a327d93c4af78d5c7e1d3be5e157 |
| ParticipantIDs | doaj_primary_oai_doaj_org_article_0a33a327d93c4af78d5c7e1d3be5e157 crossref_primary_10_1177_17483026251348851 |
| PublicationCentury | 2000 |
| PublicationDate | 2025-06-01 |
| PublicationDateYYYYMMDD | 2025-06-01 |
| PublicationDate_xml | – month: 06 year: 2025 text: 2025-06-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationTitle | Journal of algorithms & computational technology |
| PublicationYear | 2025 |
| Publisher | SAGE Publishing |
| Publisher_xml | – name: SAGE Publishing |
| References | e_1_3_2_26_2 e_1_3_2_49_2 e_1_3_2_28_2 e_1_3_2_20_2 e_1_3_2_22_2 e_1_3_2_45_2 e_1_3_2_24_2 e_1_3_2_47_2 e_1_3_2_9_2 e_1_3_2_16_2 e_1_3_2_37_2 e_1_3_2_7_2 Jadon A (e_1_3_2_8_2) e_1_3_2_18_2 e_1_3_2_39_2 Cao C (e_1_3_2_17_2) 2024; 13 Amalou I (e_1_3_2_35_2) e_1_3_2_10_2 e_1_3_2_31_2 e_1_3_2_52_2 e_1_3_2_5_2 e_1_3_2_12_2 e_1_3_2_33_2 e_1_3_2_3_2 e_1_3_2_14_2 Jaman GG (e_1_3_2_32_2) 2020; 44 e_1_3_2_50_2 He K (e_1_3_2_43_2) e_1_3_2_27_2 e_1_3_2_48_2 e_1_3_2_29_2 e_1_3_2_40_2 e_1_3_2_21_2 e_1_3_2_42_2 e_1_3_2_23_2 e_1_3_2_44_2 e_1_3_2_46_2 Zhou Z (e_1_3_2_51_2) e_1_3_2_15_2 e_1_3_2_38_2 e_1_3_2_6_2 e_1_3_2_30_2 e_1_3_2_53_2 e_1_3_2_34_2 e_1_3_2_4_2 e_1_3_2_13_2 e_1_3_2_36_2 e_1_3_2_2_2 Yadav DK (e_1_3_2_11_2) Szegedy C (e_1_3_2_41_2) Babu BM (e_1_3_2_19_2) Saini KP (e_1_3_2_25_2) 2024; 30 |
| References_xml | – ident: e_1_3_2_16_2 doi: 10.1016/j.jrmge.2023.06.015 – volume-title: Deep learning-based classification of neurodegenerative diseases using gait dataset: A comparative studyIn Proceedings of the 2023 International Conference on Robotics, Control and Vision Engineering ident: e_1_3_2_51_2 – ident: e_1_3_2_22_2 doi: 10.1186/s13677-023-00560-1 – ident: e_1_3_2_38_2 doi: 10.3390/math11030590 – volume-title: A Comprehensive Survey of Regression-Based Loss Functions for Time Series ForecastingIn International Conference on Data Management, Analytics & Innovation ident: e_1_3_2_8_2 – ident: e_1_3_2_26_2 doi: 10.3389/fdata.2023.1282541 – ident: e_1_3_2_50_2 doi: 10.1016/j.ins.2023.119147 – ident: e_1_3_2_18_2 doi: 10.1002/oca.3122 – volume: 44 start-page: 47 year: 2020 ident: e_1_3_2_32_2 article-title: Convolutional neural networks for time series data processing applicable to sEMG controlled hand prosthesis publication-title: Technische Mechanik-European Journal of Engineering Mechanics – ident: e_1_3_2_9_2 doi: 10.1016/j.jeconom.2022.12.008 – volume-title: CNN-LSTM architectures for non-stationary time series: decomposition approachIn 2024 International Conference on Global Aeronautical Engineering and Satellite Technology (GAST) ident: e_1_3_2_35_2 – ident: e_1_3_2_44_2 doi: 10.1109/ACCESS.2018.2814605 – volume: 13 start-page: 2834 year: 2024 ident: e_1_3_2_17_2 article-title: A multivariate time series prediction method based on convolution-residual gated recurrent neural network and double-layer attention publication-title: Electronics (Basel) – ident: e_1_3_2_53_2 doi: 10.1007/s10489-024-05649-x – ident: e_1_3_2_28_2 doi: 10.1007/s12145-024-01414-3 – ident: e_1_3_2_34_2 doi: 10.1007/s12541-024-01069-6 – ident: e_1_3_2_20_2 doi: 10.1080/17499518.2022.2138918 – ident: e_1_3_2_48_2 doi: 10.1007/s12517-021-08259-w – ident: e_1_3_2_40_2 doi: 10.1007/s11269-022-03419-3 – ident: e_1_3_2_46_2 doi: 10.1007/s11063-022-10929-z – volume-title: Deep residual learning for image recognitionIn Proceedings of the IEEE conference on computer vision and pattern recognition ident: e_1_3_2_43_2 – ident: e_1_3_2_31_2 doi: 10.1063/5.0172297 – ident: e_1_3_2_42_2 doi: 10.1007/s10618-020-00710-y – ident: e_1_3_2_13_2 doi: 10.7717/peerj-cs.1795 – ident: e_1_3_2_52_2 doi: 10.3390/s22010157 – ident: e_1_3_2_29_2 doi: 10.1007/s41066-023-00444-4 – volume-title: Autoregressive Integrated Moving Average Model for Time Series AnalysisInternational Conference on Optimization Computing and Wireless Communication (ICOCWC) ident: e_1_3_2_11_2 – ident: e_1_3_2_24_2 doi: 10.1007/s10618-020-00710-y – ident: e_1_3_2_2_2 doi: 10.1201/9781003254515-8 – ident: e_1_3_2_12_2 doi: 10.1080/00273171.2023.2214787 – volume-title: Going deeper with convolutionsIn Proceedings of the IEEE conference on computer vision and pattern recognition ident: e_1_3_2_41_2 – ident: e_1_3_2_6_2 doi: 10.1145/3649448 – ident: e_1_3_2_5_2 doi: 10.1007/s42979-020-00180-5 – ident: e_1_3_2_33_2 doi: 10.3390/en17091998 – ident: e_1_3_2_27_2 doi: 10.1007/s00521-024-09962-x – ident: e_1_3_2_10_2 doi: 10.1016/j.jeconom.2023.105544 – ident: e_1_3_2_39_2 doi: 10.1016/j.compag.2023.107705 – ident: e_1_3_2_23_2 doi: 10.3390/a17020076 – volume-title: A novel time series classification for multivariate data using improved deep belief-recurrent neural network with optimal dynamic time warpingIn MATEC Web of Conferences ident: e_1_3_2_19_2 – ident: e_1_3_2_14_2 doi: 10.1109/ACCESS.2024.3369891 – ident: e_1_3_2_45_2 doi: 10.1609/aaai.v31i1.11231 – ident: e_1_3_2_7_2 doi: 10.1016/j.neucom.2021.02.046 – ident: e_1_3_2_21_2 doi: 10.1089/big.2022.0155 – volume: 30 start-page: 8760 year: 2024 ident: e_1_3_2_25_2 article-title: A comparison between long short-term memory and prophet for time series analysis and forecasting technique publication-title: Educational Administration: Theory and Practice – ident: e_1_3_2_3_2 doi: 10.1109/ACCESS.2021.3091162 – ident: e_1_3_2_36_2 doi: 10.1016/j.procs.2024.03.007 – ident: e_1_3_2_37_2 doi: 10.1016/j.asoc.2022.109945 – ident: e_1_3_2_47_2 doi: 10.1109/ACCESS.2020.3023273 – ident: e_1_3_2_15_2 doi: 10.1016/j.ins.2023.119951 – ident: e_1_3_2_49_2 doi: 10.1007/s11042-021-11885-x – ident: e_1_3_2_4_2 doi: 10.1007/s10618-021-00745-9 – ident: e_1_3_2_30_2 doi: 10.1016/j.eswa.2023.121177 |
| SSID | ssj0000327813 |
| Score | 2.3068964 |
| Snippet | Accurate time-series classification (TSC) remains a fundamental challenge in deep learning due to the complexity and variability of temporal patterns. While... |
| SourceID | doaj crossref |
| SourceType | Open Website Index Database |
| Title | Hybrid deep learning framework for robust time-series classification: Integrating inception modules with residual networks |
| URI | https://doaj.org/article/0a33a327d93c4af78d5c7e1d3be5e157 |
| Volume | 19 |
| WOSCitedRecordID | wos001505375500001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVAON databaseName: Directory of Open Access Journals (DOAJ) customDbUrl: eissn: 1748-3026 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0000327813 issn: 1748-3018 databaseCode: DOA dateStart: 20070101 isFulltext: true titleUrlDefault: https://www.doaj.org/ providerName: Directory of Open Access Journals – providerCode: PRVPQU databaseName: ProQuest Central customDbUrl: eissn: 1748-3026 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0000327813 issn: 1748-3018 databaseCode: BENPR dateStart: 20160301 isFulltext: true titleUrlDefault: https://www.proquest.com/central providerName: ProQuest – providerCode: PRVPQU databaseName: Publicly Available Content Database customDbUrl: eissn: 1748-3026 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0000327813 issn: 1748-3018 databaseCode: PIMPY dateStart: 20160301 isFulltext: true titleUrlDefault: http://search.proquest.com/publiccontent providerName: ProQuest – providerCode: PRVPQU databaseName: Research Library customDbUrl: eissn: 1748-3026 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0000327813 issn: 1748-3018 databaseCode: M2O dateStart: 20160301 isFulltext: true titleUrlDefault: https://search.proquest.com/pqrl providerName: ProQuest |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV3NS8MwFA8yPehB_MT5MXLwJASbplkSbyrKdtjYQWGeSj5loN1YO0H_epO0GxMPXjy2JKW89-j7vfS93w-Ay1RZ7vOgQ766yFDmITpSNsNISymUTDRVzkWxCTYc8vFYjNakvkJPWE0PXBvuOpGESJIyI4jOpGPcUM0sNkRZajGNc-QJE2vFVPwG-x0ck-Y3ZmBY8sg7UF15uI-JD1qKfySiNb7-mFge98Bugwjhbf0m-2DDFgdgZ7CiUy0PwVfvMwxWQWPtDDY6D6_QLfuqoAeecD5Vi7KCQSsehbCyJdQBGYdWoGj9G9hvqCHC5knR9LPA96lZvPnV4UQW-uI7TmfBom4PL4_A8-PD030PNaIJSJOEV8gqkZlESSykpJQJ12WsS4WmhlDZtTooVVGmM-6vDLU61QxjIZzEqWRacXIMWsW0sCcAMmeUCXxY2vgCmjvFkiwzXPmyllgqRRtcLS2Yz2pujBw39OG_zN0Gd8HGq4WB1jre8M7OG2fnfzn79D8ecga20yDiG49SzkGrmi_sBdjSH9WknHdiHHXA5qg_GL18AxXr0Po |
| linkProvider | Directory of Open Access Journals |
| 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=Hybrid+deep+learning+framework+for+robust+time-series+classification%3A+Integrating+inception+modules+with+residual+networks&rft.jtitle=Journal+of+algorithms+%26+computational+technology&rft.au=Duong+Thi+Kim+Chi&rft.au=Nguyen+Thi+Mai+Trang&rft.au=Tran+Ba+Minh+Son&rft.au=Nguyen+Ngoc+Thao&rft.date=2025-06-01&rft.pub=SAGE+Publishing&rft.eissn=1748-3026&rft.volume=19&rft_id=info:doi/10.1177%2F17483026251348851&rft.externalDBID=DOA&rft.externalDocID=oai_doaj_org_article_0a33a327d93c4af78d5c7e1d3be5e157 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1748-3018&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1748-3018&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1748-3018&client=summon |