Feature Optimization and Stacked Ensemble Learning for Parkinson's Disease Classification Using Speech Analysis
Parkinson's Disease (PD) is the second most common neurodegenerative disorder, whose symptoms worsen over time, making early diagnosis a challenging task. Changes in speech have been identified as an early symptom of PD identification. However, medical datasets often have a small sample size, w...
Saved in:
| Published in: | Technical review - IETE Vol. 42; no. 5; pp. 632 - 650 |
|---|---|
| Main Authors: | , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Taylor & Francis
03.09.2025
|
| Subjects: | |
| ISSN: | 0256-4602, 0974-5971 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Parkinson's Disease (PD) is the second most common neurodegenerative disorder, whose symptoms worsen over time, making early diagnosis a challenging task. Changes in speech have been identified as an early symptom of PD identification. However, medical datasets often have a small sample size, while speech signal analysis generates high-dimensional data. Therefore, rigorous feature selection is essential for obtaining the best set of PD characteristics. This paper proposes a hybrid filter-wrapper feature selection approach for PD classification using a publicly available speech dataset (188 PD, 64 healthy subjects). Maximum Relevancy Minimum Redundancy (mRMR) and Relief algorithms are used to select top-ranked features, followed by the Modified Whale Optimization Algorithm (mWOA) to refine the selection for obtaining an optimized feature subset. The class imbalance issue is addressed using SMOTE. A stacked ensemble model is developed, integrating base learners, Decision Tree, Support Vector Machine, Naïve Bayes, k-Nearest Neighbour, and deep networks like shallow and deep with hyperparameters tuned via a grid search mechanism. The proposed approach is evaluated against state-of-the-art methods based on accuracy, precision, recall, and F1-score. Results demonstrate that hybrid feature selection and hyperparameter tuning reduce computational burden while improving classification accuracy, making it a promising framework for PD detection from speech data. |
|---|---|
| AbstractList | Parkinson's Disease (PD) is the second most common neurodegenerative disorder, whose symptoms worsen over time, making early diagnosis a challenging task. Changes in speech have been identified as an early symptom of PD identification. However, medical datasets often have a small sample size, while speech signal analysis generates high-dimensional data. Therefore, rigorous feature selection is essential for obtaining the best set of PD characteristics. This paper proposes a hybrid filter-wrapper feature selection approach for PD classification using a publicly available speech dataset (188 PD, 64 healthy subjects). Maximum Relevancy Minimum Redundancy (mRMR) and Relief algorithms are used to select top-ranked features, followed by the Modified Whale Optimization Algorithm (mWOA) to refine the selection for obtaining an optimized feature subset. The class imbalance issue is addressed using SMOTE. A stacked ensemble model is developed, integrating base learners, Decision Tree, Support Vector Machine, Naïve Bayes, k-Nearest Neighbour, and deep networks like shallow and deep with hyperparameters tuned via a grid search mechanism. The proposed approach is evaluated against state-of-the-art methods based on accuracy, precision, recall, and F1-score. Results demonstrate that hybrid feature selection and hyperparameter tuning reduce computational burden while improving classification accuracy, making it a promising framework for PD detection from speech data. |
| Author | Agrawal, Sneha Sahu, Satya Prakash |
| Author_xml | – sequence: 1 givenname: Sneha surname: Agrawal fullname: Agrawal, Sneha organization: National Institute of Technology Raipur – sequence: 2 givenname: Satya Prakash surname: Sahu fullname: Sahu, Satya Prakash organization: National Institute of Technology Raipur |
| BookMark | eNp9kE9LAzEQxYNUsK1-BCE3T1uTbLJ_bpbaqlCoUHsOs7sTjd3NlmRF6qd3l9arp3kzvPdgfhMycq1DQm45m3GWsXsmVCITJmaiV7N-YRmPL8iY5amMVJ7yUa_7czSYrsgkhE_GEikUH5N2hdB9eaSbQ2cb-wOdbR0FV9FtB-UeK7p0AZuiRrpG8M66d2paT1_B760LrbsL9NEGhIB0UUMI1tjyVLILg3l7QCw_6NxBfQw2XJNLA3XAm_Ockt1q-bZ4jtabp5fFfB2VImFdVKSQoSilYQUIzLlRWGZFxtKcJ2kmijSXlcxVyk0BHEAmCvO4yJVi0oBMq3hK1Km39G0IHo0-eNuAP2rO9EBN_1HTAzV9ptbnHk456_o3G_hufV3pDo51640HV9qg4_8rfgHMFXcq |
| Cites_doi | 10.1016/j.neucom.2015.07.138 10.1155/2017/6209703 10.32604/csse.2022.022739 10.1007/3-540-57868-4_57 10.1016/j.bbe.2022.07.002 10.3390/math12101575 10.1038/npre.2008.2298.1 10.1016/j.engappai.2022.105151 10.1016/j.compbiomed.2018.09.008 10.1109/SIBGRAPI.2016.054 10.1016/j.jocn.2019.12.029 10.1016/j.neucom.2016.03.101 10.1109/SPMB.2018.8615607 10.1109/ACCESS.2025.3542160 10.1016/j.bspc.2022.104553 10.1007/s00521-018-3689-5 10.1007/s11761-023-00372-w 10.1109/LSP.2005.860538 10.1613/jair.953 10.1109/ACCESS.2019.2936564 10.1016/j.ins.2017.05.013 10.1016/j.bbe.2020.01.003 10.1016/j.advengsoft.2016.01.008 10.1016/B978-1-55860-247-2.50037-1 10.18517/ijaseit.11.1.12202 10.1214/07-AOS537 10.1007/s13369-021-06544-0 10.1007/978-1-4757-2440-0 10.1136/jnnp.2007.131045 10.1109/LifeTech52111.2021.9391925 10.1109/TBME.2006.871883 10.1080/02564602.2022.2121772 10.1109/LSENS.2020.2994938 10.1109/TSP.2011.2143711 10.1109/EMBC.2016.7590787 10.1016/j.cose.2024.104288 10.1002/mds.26693 10.1080/00220670209598786 10.1016/j.asoc.2018.10.022 10.1109/IJCNN.2019.8851995 10.1109/JBHI.2013.2245674 10.1109/ICASSP49660.2025.10888319 10.1080/03772063.2018.1531730 |
| ContentType | Journal Article |
| Copyright | 2025 IETE 2025 |
| Copyright_xml | – notice: 2025 IETE 2025 |
| DBID | AAYXX CITATION |
| DOI | 10.1080/02564602.2025.2560813 |
| DatabaseName | CrossRef |
| DatabaseTitle | CrossRef |
| DatabaseTitleList | |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Engineering |
| EISSN | 0974-5971 |
| EndPage | 650 |
| ExternalDocumentID | 10_1080_02564602_2025_2560813 2560813 |
| Genre | Research Article |
| GroupedDBID | .DC 0BK 0R~ 29Q 2WC 30N 4.4 5GY AAGDL AAHIA AAJMT AALDU AAMIU AAPUL AAQRR ABCCY ABFIM ABJNI ABLIJ ABPAQ ABPEM ABTAI ABXUL ABXYU ACGFS ACTIO ADCVX ADGTB AEISY AENEX AEYOC AFRVT AGDLA AHDZW AIDUJ AIJEM AIYEW AKBVH AKOOK ALMA_UNASSIGNED_HOLDINGS ALQZU AQRUH AQTUD AVBZW AWYRJ BLEHA CCCUG DKSSO DU5 E3Z EBS GTTXZ H13 HZ~ IPNFZ KYCEM LJTGL M4Z O9- P2P RIG RNANH RNS ROSJB RTWRZ SC5 SNACF TAJZE TASJS TBQAZ TDBHL TEN TFL TFT TFW TR2 TTHFI TUROJ ZGOLN AAYXX CITATION DGEBU |
| ID | FETCH-LOGICAL-c260t-b7a8e2c4f0ba2e91f5ec8b807916782b794d49571fba1aa465e93b95504fa47d3 |
| IEDL.DBID | TFW |
| ISICitedReferencesCount | 0 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001586563200001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 0256-4602 |
| IngestDate | Sat Nov 29 06:57:04 EST 2025 Sat Nov 01 10:41:24 EDT 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 5 |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c260t-b7a8e2c4f0ba2e91f5ec8b807916782b794d49571fba1aa465e93b95504fa47d3 |
| PageCount | 19 |
| ParticipantIDs | informaworld_taylorfrancis_310_1080_02564602_2025_2560813 crossref_primary_10_1080_02564602_2025_2560813 |
| PublicationCentury | 2000 |
| PublicationDate | 2025-09-03 |
| PublicationDateYYYYMMDD | 2025-09-03 |
| PublicationDate_xml | – month: 09 year: 2025 text: 2025-09-03 day: 03 |
| PublicationDecade | 2020 |
| PublicationTitle | Technical review - IETE |
| PublicationYear | 2025 |
| Publisher | Taylor & Francis |
| Publisher_xml | – name: Taylor & Francis |
| References | e_1_3_3_52_1 e_1_3_3_50_1 Sharma A. (e_1_3_3_8_1) 2014; 4 Bchir O. (e_1_3_3_14_1) 2020; 11 Alemami Y. (e_1_3_3_9_1) 2014; 10 e_1_3_3_18_1 Ghanad N. K. (e_1_3_3_37_1) 2015; 4 e_1_3_3_39_1 Rouzbahani H. K. (e_1_3_3_13_1) 2011; 2 e_1_3_3_16_1 e_1_3_3_35_1 e_1_3_3_10_1 e_1_3_3_33_1 e_1_3_3_12_1 e_1_3_3_31_1 e_1_3_3_40_1 e_1_3_3_7_1 e_1_3_3_29_1 e_1_3_3_25_1 e_1_3_3_48_1 e_1_3_3_27_1 e_1_3_3_46_1 Caliskan A. (e_1_3_3_20_1) 2017; 17 e_1_3_3_3_1 e_1_3_3_21_1 e_1_3_3_44_1 e_1_3_3_5_1 e_1_3_3_23_1 e_1_3_3_42_1 e_1_3_3_30_1 e_1_3_3_51_1 De Rijk M. D. (e_1_3_3_2_1) 2000; 54 e_1_3_3_17_1 e_1_3_3_19_1 e_1_3_3_38_1 e_1_3_3_15_1 e_1_3_3_36_1 e_1_3_3_34_1 e_1_3_3_11_1 e_1_3_3_32_1 e_1_3_3_41_1 e_1_3_3_6_1 e_1_3_3_28_1 e_1_3_3_24_1 e_1_3_3_49_1 e_1_3_3_26_1 e_1_3_3_47_1 e_1_3_3_45_1 e_1_3_3_4_1 e_1_3_3_22_1 e_1_3_3_43_1 |
| References_xml | – ident: e_1_3_3_5_1 doi: 10.1016/j.neucom.2015.07.138 – ident: e_1_3_3_21_1 doi: 10.1155/2017/6209703 – ident: e_1_3_3_47_1 doi: 10.32604/csse.2022.022739 – ident: e_1_3_3_34_1 doi: 10.1007/3-540-57868-4_57 – ident: e_1_3_3_35_1 doi: 10.1016/j.bbe.2022.07.002 – ident: e_1_3_3_17_1 doi: 10.3390/math12101575 – ident: e_1_3_3_11_1 doi: 10.1038/npre.2008.2298.1 – volume: 4 start-page: 2278 issue: 3 year: 2014 ident: e_1_3_3_8_1 article-title: Automatic recognition of Parkinson’s disease via artificial neural network and support vector machine publication-title: Int. J. Innov. Technol. Explor. Eng. (IJITEE) – ident: e_1_3_3_41_1 doi: 10.1016/j.engappai.2022.105151 – ident: e_1_3_3_7_1 doi: 10.1016/j.compbiomed.2018.09.008 – ident: e_1_3_3_25_1 doi: 10.1109/SIBGRAPI.2016.054 – ident: e_1_3_3_4_1 doi: 10.1016/j.jocn.2019.12.029 – ident: e_1_3_3_49_1 doi: 10.1016/j.neucom.2016.03.101 – ident: e_1_3_3_19_1 doi: 10.1109/SPMB.2018.8615607 – volume: 11 start-page: 413 issue: 4 year: 2020 ident: e_1_3_3_14_1 article-title: Parkinson’s disease classification using Gaussian mixture models with relevance feature weights on vocal feature sets publication-title: Int. J. Adv. Comput. Sci. Appl. – ident: e_1_3_3_50_1 doi: 10.1109/ACCESS.2025.3542160 – volume: 17 start-page: 3311 issue: 2 year: 2017 ident: e_1_3_3_20_1 article-title: Diagnosis of the Parkinson disease by using deep neural network classifier publication-title: IU-J. Electr. Electron. Eng. – ident: e_1_3_3_42_1 doi: 10.1016/j.bspc.2022.104553 – ident: e_1_3_3_23_1 doi: 10.1007/s00521-018-3689-5 – ident: e_1_3_3_15_1 doi: 10.1007/s11761-023-00372-w – ident: e_1_3_3_27_1 doi: 10.1109/LSP.2005.860538 – ident: e_1_3_3_31_1 doi: 10.1613/jair.953 – ident: e_1_3_3_18_1 doi: 10.1109/ACCESS.2019.2936564 – ident: e_1_3_3_32_1 doi: 10.1016/j.ins.2017.05.013 – ident: e_1_3_3_46_1 doi: 10.1016/j.bbe.2020.01.003 – ident: e_1_3_3_48_1 doi: 10.1016/j.advengsoft.2016.01.008 – ident: e_1_3_3_33_1 doi: 10.1016/B978-1-55860-247-2.50037-1 – ident: e_1_3_3_16_1 doi: 10.18517/ijaseit.11.1.12202 – ident: e_1_3_3_39_1 doi: 10.1214/07-AOS537 – ident: e_1_3_3_45_1 doi: 10.1007/s13369-021-06544-0 – volume: 54 start-page: S21 issue: 11 year: 2000 ident: e_1_3_3_2_1 article-title: Prevalence of Parkinson's disease in Europe: A collaborative study of population-based cohorts. Neurologic diseases in the elderly research group publication-title: Neurology – ident: e_1_3_3_38_1 doi: 10.1007/978-1-4757-2440-0 – volume: 10 start-page: 44 issue: 10 year: 2014 ident: e_1_3_3_9_1 article-title: Detection of Parkinson disease through voice signal features publication-title: J. Amer. Sci. – ident: e_1_3_3_3_1 doi: 10.1136/jnnp.2007.131045 – ident: e_1_3_3_44_1 doi: 10.1109/LifeTech52111.2021.9391925 – ident: e_1_3_3_28_1 doi: 10.1109/TBME.2006.871883 – ident: e_1_3_3_43_1 doi: 10.1080/02564602.2022.2121772 – ident: e_1_3_3_29_1 – volume: 2 start-page: 12 issue: 3 year: 2011 ident: e_1_3_3_13_1 article-title: Diagnosis of Parkinson's disease in human using voice signals publication-title: Basic. Clin. Neurosci. – ident: e_1_3_3_24_1 doi: 10.1109/LSENS.2020.2994938 – ident: e_1_3_3_30_1 doi: 10.1109/TSP.2011.2143711 – ident: e_1_3_3_26_1 doi: 10.1109/EMBC.2016.7590787 – ident: e_1_3_3_51_1 doi: 10.1016/j.cose.2024.104288 – ident: e_1_3_3_10_1 doi: 10.1002/mds.26693 – volume: 4 start-page: 119 issue: 4 year: 2015 ident: e_1_3_3_37_1 article-title: Combination of PSO algorithm and naive Bayesian classification for Parkinson disease diagnosis publication-title: Adv. Comput. Sci.: Int. J. – ident: e_1_3_3_36_1 doi: 10.1080/00220670209598786 – ident: e_1_3_3_6_1 doi: 10.1016/j.asoc.2018.10.022 – ident: e_1_3_3_22_1 doi: 10.1109/IJCNN.2019.8851995 – ident: e_1_3_3_12_1 doi: 10.1109/JBHI.2013.2245674 – ident: e_1_3_3_52_1 doi: 10.1109/ICASSP49660.2025.10888319 – ident: e_1_3_3_40_1 doi: 10.1080/03772063.2018.1531730 |
| SSID | ssj0064251 |
| Score | 2.3444655 |
| Snippet | Parkinson's Disease (PD) is the second most common neurodegenerative disorder, whose symptoms worsen over time, making early diagnosis a challenging task.... |
| SourceID | crossref informaworld |
| SourceType | Index Database Publisher |
| StartPage | 632 |
| SubjectTerms | Grid search optimization Hybrid filter-wrapper Hyperparameter tuning Maximum relevancy minimum redundancy Parkinson's disease ReliefF Synthetic Minority Oversampling Technique (SMOTE) Whale optimization algorithm |
| Title | Feature Optimization and Stacked Ensemble Learning for Parkinson's Disease Classification Using Speech Analysis |
| URI | https://www.tandfonline.com/doi/abs/10.1080/02564602.2025.2560813 |
| Volume | 42 |
| WOSCitedRecordID | wos001586563200001&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: PRVAWR databaseName: Taylor & Francis Journals Complete customDbUrl: eissn: 0974-5971 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0064251 issn: 0256-4602 databaseCode: TFW dateStart: 19840101 isFulltext: true titleUrlDefault: https://www.tandfonline.com providerName: Taylor & Francis |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV05T8MwFLZQxQADN6Jc8oDE5JLDie0RQSsGVJAo0C3yCUg0rZqU34_tOFI7wAJjhmdbvt73nO99D4CLjBOVSyYRZZIhrCKDqBEGmTjSEbfRW-afsl_uyXBIx2P2GNiEVaBVuhjaNEIR_q52h5uLqmXEXTk3jXOfRpVkPeezqa9ba5G9I_WNBq_tXWzBtS_A6CyQM2lzeH5qZcU7rWiXLnmdwfY_jHcHbAXICa-bPbIL1nS5BzaXhAj3wdQhwcVcwwd7g0xCaia0HUILRu05V7BfVnoiPjUMgqxv0A4aupxpnz52WcHb5lcP9GU2HQGpacRzEuDTTGv5DlsJlAPwPOiPbu5QKMWApA14aiQIpzqR2ESCJ5rFJtOSChoRiy4txhD2VCsbapHYCB5zjvNMs1QwG_5gwzFR6SHolNNSHwEolAUNmDKVMolzY3jEU5xrwhOlc5GSLui1S1DMGsWNIm6FTMNMFm4mizCTXcCWF6qo_VOHaeqSFOmvtsd_sD0BG-7TE87SU9Cp5wt9BtblV_1Rzc_9PvwGS5bbRw |
| linkProvider | Taylor & Francis |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV07T8MwELZQQQIG3ojy9IDE5JI0TmyPCKiKKGWgQLfIT0CiadWm_H5sJ5HSARaYo7Ot8-O-u9x9B8B5zIlKJJOIMskQVoFB1AiDTBjogFvvLfah7Jce6ffpcMjqtTAurdL50KYgivBvtbvcLhhdpcRdOjuNE19H1Y5bzmhT17h22XWncw7YoPNavcYWXvsWjE4EOZmqiuenYRbs0wJ7ac3udDb_Y8VbYKNEnfCqOCbbYElnO2C9xkW4C8YODM6nGj7aR2RUVmdCOyO0eNRedQVvs5keiU8NS07WN2hXDV3ZtK8gu5jBm-JvD_SdNl0OUjGIT0uATxOt5TusWFD2wHPndnDdRWU3BiStz5MjQTjVbYlNIHhbs9DEWlJBA2IBpoUZwl5sZb0tEhrBQ85xEmsWCWY9IGw4JiraB41snOkDAIWyuAFTpiImcWIMD3iEE014W-lERKQJWtUepJOCdCMNKy7TUpOp02RaarIJWH2n0txHO0zRmiSNfpU9_IPsGVjtDh56ae-uf38E1twnn38WHYNGPp3rE7Aiv_KP2fTUH8pvKuHfag |
| linkToPdf | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV07T8MwELZQQQgG3ojy9IDElJKHE8cjoq1AVKUSBbpFfgISTasm5fdjO47UDrDAHJ1tne2775y77wC4jCkWCSfcSwknHhK-8lLFlKcCX_pUR2-xfcp-6eF-Px2NyMBlExYurdLE0KoiirC22lzuqVB1Rty1cdMosWVUYdwyPjs1fWtXNXSOzcEedl9rY6zRte3AaEQ8I1MX8fw0zJJ7WiIvXXA73e1_WPAO2HKYE95Uh2QXrMh8D2wuMBHug4mBgvOZhI_ahIxdbSbUE0KNRvVFF7CTF3LMPiV0jKxvUC8amqJpWz92VcB29a8H2j6bJgOpGsQmJcCnqZT8HdYcKAfgudsZ3t55rheDx3XEU3oM01SGHCmf0VCSQMWSpyz1sYaXGmQwfa2FjrVwoBgNKEVJLEnEiI5_kKIIi-gQNPJJLo8AZEKjBpQSERGOEqWoTyOUSExDIRMW4SZo1VuQTSvKjSyomUydJjOjycxpsgnI4kZlpX3rUFVjkiz6Vfb4D7IXYH3Q7ma9-_7DCdgwX2zyWXQKGuVsLs_AGv8qP4rZuT2S36Fg3hw |
| 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=Feature+Optimization+and+Stacked+Ensemble+Learning+for+Parkinson%E2%80%99s+Disease+Classification+Using+Speech+Analysis&rft.jtitle=Technical+review+-+IETE&rft.au=Agrawal%2C+Sneha&rft.au=Sahu%2C+Satya+Prakash&rft.date=2025-09-03&rft.issn=0256-4602&rft.eissn=0974-5971&rft.volume=42&rft.issue=5&rft.spage=632&rft.epage=650&rft_id=info:doi/10.1080%2F02564602.2025.2560813&rft.externalDBID=n%2Fa&rft.externalDocID=10_1080_02564602_2025_2560813 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0256-4602&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0256-4602&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0256-4602&client=summon |