Assessment of input variables determination on the SVM model performance using PCA, Gamma test, and forward selection techniques for monthly stream flow prediction

► Role of three input selection is evaluated on SVM performance for prediction of monthly stream flow. ► Comparison among the developed SVM and ANN models is carried out. ► A new statistic is introduced to evaluate the performance of intelligent models. In the research, the role of three input selec...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of hydrology (Amsterdam) 2011-05, Vol.401 (3), p.177-189
Hauptverfasser: Noori, R., Karbassi, A.R., Moghaddamnia, A., Han, D., Zokaei-Ashtiani, M.H., Farokhnia, A., Gousheh, M. Ghafari
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 189
container_issue 3
container_start_page 177
container_title Journal of hydrology (Amsterdam)
container_volume 401
creator Noori, R.
Karbassi, A.R.
Moghaddamnia, A.
Han, D.
Zokaei-Ashtiani, M.H.
Farokhnia, A.
Gousheh, M. Ghafari
description ► Role of three input selection is evaluated on SVM performance for prediction of monthly stream flow. ► Comparison among the developed SVM and ANN models is carried out. ► A new statistic is introduced to evaluate the performance of intelligent models. In the research, the role of three input selection techniques is evaluated on support vector machine (SVM) performance for prediction of monthly stream flow. First, a SVM model is adapted to predict the next monthly flow as a function of 18 input variables including monthly rainfall ( R), discharge ( Q), sun radiation (Rad), and temperature {as minimum ( T min), maximum ( T max) and average ( T ave)} with three temporal delays belong to t, t-1, and t-2. Subsequently, principal component analysis (PCA), Gamma test (GT), and forward selection (FS) techniques are used to reduce the number of input variables. Upon reducing 18 input variables to 5 (using PCA and GT) and 7 (using FS techniques), they are then fed to SVM model. In addition, a proper artificial neural network (ANN) model based on PCA technique is developed (PCA-ANN). Then, comparison among the developed SVM models (PCA-SVM and GT-SVM) and PCA-ANN model is carried out. Furthermore, the imperfections of the discrepancy ratio (DR) statistic are remedied and an appropriate DR statistic is developed. Finally, the error distribution during testing step of selected models (PCA-SVM, GT-SVM, and PCA-ANN) is computed using the developed DR statistic. Results indicated that preprocessing the input variables by means of PCA and GT techniques has improved the SVM model operation and the developed models (PCA-SVM and GT-SVM) are considerably better than original SVM model. Besides, PCA-SVM is superior to GT-SVM and PCA-ANN models. Determination coefficient ( R 2) for PCA-SVM model was equal to 0.92 and 0.88 in the training and testing steps, respectively.
doi_str_mv 10.1016/j.jhydrol.2011.02.021
format Article
fullrecord <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_876226007</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><els_id>S0022169411001363</els_id><sourcerecordid>876226007</sourcerecordid><originalsourceid>FETCH-LOGICAL-c371t-1efd862a392baa8be75b6b5ca1b2da5994e156dadcbdf19b500f62bd345468c63</originalsourceid><addsrcrecordid>eNqFUU2P0zAUjBBIlIWfgOQL4rIptpM4yQlV1bIgLQKJj6v1bL9QV44dbHdX_T38UdxtxRVrJB88b-Z5pqpeM7pmlIl3-_V-dzQxuDWnjK0pL2BPqhUb-rHmPe2fVitKOa-ZGNvn1YuU9rScpmlX1Z9NSpjSjD6TMBHrl0Mm9xAtKIeJGMwYZ-sh2-BJQd4h-fbzM5mDQUcWjFOIM3iN5JCs_0W-bjfX5BbmGUjGlK8JeEMK5wGiIQkd6keljHrn7e9DsSiPRc3nnTuSlCPCTCYXHsgS0dhH9svq2QQu4avLfVX9-HDzffuxvvty-2m7uat107NcM5zMIDg0I1cAg8K-U0J1GpjiBrpxbJF1woDRykxsVB2lk-DKNG3XikGL5qp6e9ZdYjitluVsk0bnwGM4JDn0gnNBaV-Y3ZmpY0gp4iSXaGeIR8moPHUi9_LSiTx1IikvYGXuzcUBkgY3xZKcTf-GeUt7wTpaeO_PPCzfvbcYZdIWS8rGxpKgNMH-x-kvGD6pag</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>876226007</pqid></control><display><type>article</type><title>Assessment of input variables determination on the SVM model performance using PCA, Gamma test, and forward selection techniques for monthly stream flow prediction</title><source>Elsevier ScienceDirect Journals Complete</source><creator>Noori, R. ; Karbassi, A.R. ; Moghaddamnia, A. ; Han, D. ; Zokaei-Ashtiani, M.H. ; Farokhnia, A. ; Gousheh, M. Ghafari</creator><creatorcontrib>Noori, R. ; Karbassi, A.R. ; Moghaddamnia, A. ; Han, D. ; Zokaei-Ashtiani, M.H. ; Farokhnia, A. ; Gousheh, M. Ghafari</creatorcontrib><description>► Role of three input selection is evaluated on SVM performance for prediction of monthly stream flow. ► Comparison among the developed SVM and ANN models is carried out. ► A new statistic is introduced to evaluate the performance of intelligent models. In the research, the role of three input selection techniques is evaluated on support vector machine (SVM) performance for prediction of monthly stream flow. First, a SVM model is adapted to predict the next monthly flow as a function of 18 input variables including monthly rainfall ( R), discharge ( Q), sun radiation (Rad), and temperature {as minimum ( T min), maximum ( T max) and average ( T ave)} with three temporal delays belong to t, t-1, and t-2. Subsequently, principal component analysis (PCA), Gamma test (GT), and forward selection (FS) techniques are used to reduce the number of input variables. Upon reducing 18 input variables to 5 (using PCA and GT) and 7 (using FS techniques), they are then fed to SVM model. In addition, a proper artificial neural network (ANN) model based on PCA technique is developed (PCA-ANN). Then, comparison among the developed SVM models (PCA-SVM and GT-SVM) and PCA-ANN model is carried out. Furthermore, the imperfections of the discrepancy ratio (DR) statistic are remedied and an appropriate DR statistic is developed. Finally, the error distribution during testing step of selected models (PCA-SVM, GT-SVM, and PCA-ANN) is computed using the developed DR statistic. Results indicated that preprocessing the input variables by means of PCA and GT techniques has improved the SVM model operation and the developed models (PCA-SVM and GT-SVM) are considerably better than original SVM model. Besides, PCA-SVM is superior to GT-SVM and PCA-ANN models. Determination coefficient ( R 2) for PCA-SVM model was equal to 0.92 and 0.88 in the training and testing steps, respectively.</description><identifier>ISSN: 0022-1694</identifier><identifier>EISSN: 1879-2707</identifier><identifier>DOI: 10.1016/j.jhydrol.2011.02.021</identifier><identifier>CODEN: JHYDA7</identifier><language>eng</language><publisher>Kidlington: Elsevier B.V</publisher><subject>ANN ; Earth sciences ; Earth, ocean, space ; Exact sciences and technology ; Flow prediction ; Freshwater ; Hydrology. Hydrogeology ; Input selection techniques ; Learning theory ; Mathematical analysis ; Mathematical models ; Neural networks ; Performance evaluation ; Statistics ; Streams ; Support vector machines ; SVM</subject><ispartof>Journal of hydrology (Amsterdam), 2011-05, Vol.401 (3), p.177-189</ispartof><rights>2011 Elsevier B.V.</rights><rights>2015 INIST-CNRS</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c371t-1efd862a392baa8be75b6b5ca1b2da5994e156dadcbdf19b500f62bd345468c63</citedby><cites>FETCH-LOGICAL-c371t-1efd862a392baa8be75b6b5ca1b2da5994e156dadcbdf19b500f62bd345468c63</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://dx.doi.org/10.1016/j.jhydrol.2011.02.021$$EHTML$$P50$$Gelsevier$$H</linktohtml><link.rule.ids>314,780,784,3550,27924,27925,45995</link.rule.ids><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&amp;idt=24076150$$DView record in Pascal Francis$$Hfree_for_read</backlink></links><search><creatorcontrib>Noori, R.</creatorcontrib><creatorcontrib>Karbassi, A.R.</creatorcontrib><creatorcontrib>Moghaddamnia, A.</creatorcontrib><creatorcontrib>Han, D.</creatorcontrib><creatorcontrib>Zokaei-Ashtiani, M.H.</creatorcontrib><creatorcontrib>Farokhnia, A.</creatorcontrib><creatorcontrib>Gousheh, M. Ghafari</creatorcontrib><title>Assessment of input variables determination on the SVM model performance using PCA, Gamma test, and forward selection techniques for monthly stream flow prediction</title><title>Journal of hydrology (Amsterdam)</title><description>► Role of three input selection is evaluated on SVM performance for prediction of monthly stream flow. ► Comparison among the developed SVM and ANN models is carried out. ► A new statistic is introduced to evaluate the performance of intelligent models. In the research, the role of three input selection techniques is evaluated on support vector machine (SVM) performance for prediction of monthly stream flow. First, a SVM model is adapted to predict the next monthly flow as a function of 18 input variables including monthly rainfall ( R), discharge ( Q), sun radiation (Rad), and temperature {as minimum ( T min), maximum ( T max) and average ( T ave)} with three temporal delays belong to t, t-1, and t-2. Subsequently, principal component analysis (PCA), Gamma test (GT), and forward selection (FS) techniques are used to reduce the number of input variables. Upon reducing 18 input variables to 5 (using PCA and GT) and 7 (using FS techniques), they are then fed to SVM model. In addition, a proper artificial neural network (ANN) model based on PCA technique is developed (PCA-ANN). Then, comparison among the developed SVM models (PCA-SVM and GT-SVM) and PCA-ANN model is carried out. Furthermore, the imperfections of the discrepancy ratio (DR) statistic are remedied and an appropriate DR statistic is developed. Finally, the error distribution during testing step of selected models (PCA-SVM, GT-SVM, and PCA-ANN) is computed using the developed DR statistic. Results indicated that preprocessing the input variables by means of PCA and GT techniques has improved the SVM model operation and the developed models (PCA-SVM and GT-SVM) are considerably better than original SVM model. Besides, PCA-SVM is superior to GT-SVM and PCA-ANN models. Determination coefficient ( R 2) for PCA-SVM model was equal to 0.92 and 0.88 in the training and testing steps, respectively.</description><subject>ANN</subject><subject>Earth sciences</subject><subject>Earth, ocean, space</subject><subject>Exact sciences and technology</subject><subject>Flow prediction</subject><subject>Freshwater</subject><subject>Hydrology. Hydrogeology</subject><subject>Input selection techniques</subject><subject>Learning theory</subject><subject>Mathematical analysis</subject><subject>Mathematical models</subject><subject>Neural networks</subject><subject>Performance evaluation</subject><subject>Statistics</subject><subject>Streams</subject><subject>Support vector machines</subject><subject>SVM</subject><issn>0022-1694</issn><issn>1879-2707</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2011</creationdate><recordtype>article</recordtype><recordid>eNqFUU2P0zAUjBBIlIWfgOQL4rIptpM4yQlV1bIgLQKJj6v1bL9QV44dbHdX_T38UdxtxRVrJB88b-Z5pqpeM7pmlIl3-_V-dzQxuDWnjK0pL2BPqhUb-rHmPe2fVitKOa-ZGNvn1YuU9rScpmlX1Z9NSpjSjD6TMBHrl0Mm9xAtKIeJGMwYZ-sh2-BJQd4h-fbzM5mDQUcWjFOIM3iN5JCs_0W-bjfX5BbmGUjGlK8JeEMK5wGiIQkd6keljHrn7e9DsSiPRc3nnTuSlCPCTCYXHsgS0dhH9svq2QQu4avLfVX9-HDzffuxvvty-2m7uat107NcM5zMIDg0I1cAg8K-U0J1GpjiBrpxbJF1woDRykxsVB2lk-DKNG3XikGL5qp6e9ZdYjitluVsk0bnwGM4JDn0gnNBaV-Y3ZmpY0gp4iSXaGeIR8moPHUi9_LSiTx1IikvYGXuzcUBkgY3xZKcTf-GeUt7wTpaeO_PPCzfvbcYZdIWS8rGxpKgNMH-x-kvGD6pag</recordid><startdate>20110503</startdate><enddate>20110503</enddate><creator>Noori, R.</creator><creator>Karbassi, A.R.</creator><creator>Moghaddamnia, A.</creator><creator>Han, D.</creator><creator>Zokaei-Ashtiani, M.H.</creator><creator>Farokhnia, A.</creator><creator>Gousheh, M. Ghafari</creator><general>Elsevier B.V</general><general>Elsevier</general><scope>IQODW</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>8FD</scope><scope>FR3</scope><scope>KR7</scope></search><sort><creationdate>20110503</creationdate><title>Assessment of input variables determination on the SVM model performance using PCA, Gamma test, and forward selection techniques for monthly stream flow prediction</title><author>Noori, R. ; Karbassi, A.R. ; Moghaddamnia, A. ; Han, D. ; Zokaei-Ashtiani, M.H. ; Farokhnia, A. ; Gousheh, M. Ghafari</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c371t-1efd862a392baa8be75b6b5ca1b2da5994e156dadcbdf19b500f62bd345468c63</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2011</creationdate><topic>ANN</topic><topic>Earth sciences</topic><topic>Earth, ocean, space</topic><topic>Exact sciences and technology</topic><topic>Flow prediction</topic><topic>Freshwater</topic><topic>Hydrology. Hydrogeology</topic><topic>Input selection techniques</topic><topic>Learning theory</topic><topic>Mathematical analysis</topic><topic>Mathematical models</topic><topic>Neural networks</topic><topic>Performance evaluation</topic><topic>Statistics</topic><topic>Streams</topic><topic>Support vector machines</topic><topic>SVM</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Noori, R.</creatorcontrib><creatorcontrib>Karbassi, A.R.</creatorcontrib><creatorcontrib>Moghaddamnia, A.</creatorcontrib><creatorcontrib>Han, D.</creatorcontrib><creatorcontrib>Zokaei-Ashtiani, M.H.</creatorcontrib><creatorcontrib>Farokhnia, A.</creatorcontrib><creatorcontrib>Gousheh, M. Ghafari</creatorcontrib><collection>Pascal-Francis</collection><collection>CrossRef</collection><collection>Technology Research Database</collection><collection>Engineering Research Database</collection><collection>Civil Engineering Abstracts</collection><jtitle>Journal of hydrology (Amsterdam)</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Noori, R.</au><au>Karbassi, A.R.</au><au>Moghaddamnia, A.</au><au>Han, D.</au><au>Zokaei-Ashtiani, M.H.</au><au>Farokhnia, A.</au><au>Gousheh, M. Ghafari</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Assessment of input variables determination on the SVM model performance using PCA, Gamma test, and forward selection techniques for monthly stream flow prediction</atitle><jtitle>Journal of hydrology (Amsterdam)</jtitle><date>2011-05-03</date><risdate>2011</risdate><volume>401</volume><issue>3</issue><spage>177</spage><epage>189</epage><pages>177-189</pages><issn>0022-1694</issn><eissn>1879-2707</eissn><coden>JHYDA7</coden><abstract>► Role of three input selection is evaluated on SVM performance for prediction of monthly stream flow. ► Comparison among the developed SVM and ANN models is carried out. ► A new statistic is introduced to evaluate the performance of intelligent models. In the research, the role of three input selection techniques is evaluated on support vector machine (SVM) performance for prediction of monthly stream flow. First, a SVM model is adapted to predict the next monthly flow as a function of 18 input variables including monthly rainfall ( R), discharge ( Q), sun radiation (Rad), and temperature {as minimum ( T min), maximum ( T max) and average ( T ave)} with three temporal delays belong to t, t-1, and t-2. Subsequently, principal component analysis (PCA), Gamma test (GT), and forward selection (FS) techniques are used to reduce the number of input variables. Upon reducing 18 input variables to 5 (using PCA and GT) and 7 (using FS techniques), they are then fed to SVM model. In addition, a proper artificial neural network (ANN) model based on PCA technique is developed (PCA-ANN). Then, comparison among the developed SVM models (PCA-SVM and GT-SVM) and PCA-ANN model is carried out. Furthermore, the imperfections of the discrepancy ratio (DR) statistic are remedied and an appropriate DR statistic is developed. Finally, the error distribution during testing step of selected models (PCA-SVM, GT-SVM, and PCA-ANN) is computed using the developed DR statistic. Results indicated that preprocessing the input variables by means of PCA and GT techniques has improved the SVM model operation and the developed models (PCA-SVM and GT-SVM) are considerably better than original SVM model. Besides, PCA-SVM is superior to GT-SVM and PCA-ANN models. Determination coefficient ( R 2) for PCA-SVM model was equal to 0.92 and 0.88 in the training and testing steps, respectively.</abstract><cop>Kidlington</cop><pub>Elsevier B.V</pub><doi>10.1016/j.jhydrol.2011.02.021</doi><tpages>13</tpages></addata></record>
fulltext fulltext
identifier ISSN: 0022-1694
ispartof Journal of hydrology (Amsterdam), 2011-05, Vol.401 (3), p.177-189
issn 0022-1694
1879-2707
language eng
recordid cdi_proquest_miscellaneous_876226007
source Elsevier ScienceDirect Journals Complete
subjects ANN
Earth sciences
Earth, ocean, space
Exact sciences and technology
Flow prediction
Freshwater
Hydrology. Hydrogeology
Input selection techniques
Learning theory
Mathematical analysis
Mathematical models
Neural networks
Performance evaluation
Statistics
Streams
Support vector machines
SVM
title Assessment of input variables determination on the SVM model performance using PCA, Gamma test, and forward selection techniques for monthly stream flow prediction
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T13%3A06%3A13IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Assessment%20of%20input%20variables%20determination%20on%20the%20SVM%20model%20performance%20using%20PCA,%20Gamma%20test,%20and%20forward%20selection%20techniques%20for%20monthly%20stream%20flow%20prediction&rft.jtitle=Journal%20of%20hydrology%20(Amsterdam)&rft.au=Noori,%20R.&rft.date=2011-05-03&rft.volume=401&rft.issue=3&rft.spage=177&rft.epage=189&rft.pages=177-189&rft.issn=0022-1694&rft.eissn=1879-2707&rft.coden=JHYDA7&rft_id=info:doi/10.1016/j.jhydrol.2011.02.021&rft_dat=%3Cproquest_cross%3E876226007%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=876226007&rft_id=info:pmid/&rft_els_id=S0022169411001363&rfr_iscdi=true