On the Kalman filtering method in neural network training and pruning
In the use of the extended Kalman filter approach in training and pruning a feedforward neural network, one usually encounters the problems of how to set the initial condition and how to use the result obtained to prune a neural network. In this paper, some cues on the setting of the initial conditi...
Gespeichert in:
| Veröffentlicht in: | IEEE transactions on neural networks 1999-01, Vol.10 (1), p.161-166 |
|---|---|
| Hauptverfasser: | , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 166 |
|---|---|
| container_issue | 1 |
| container_start_page | 161 |
| container_title | IEEE transactions on neural networks |
| container_volume | 10 |
| creator | Sum, J. Chi-Sing Leung Young, G.H. Wing-Kay Kan |
| description | In the use of the extended Kalman filter approach in training and pruning a feedforward neural network, one usually encounters the problems of how to set the initial condition and how to use the result obtained to prune a neural network. In this paper, some cues on the setting of the initial condition are presented with a simple example illustrated. Then based on three assumptions: 1) the size of training set is large enough; 2) the training is able to converge; and 3) the trained network model is close to the actual one, an elegant equation linking the error sensitivity measure (the saliency) and the result obtained via an extended Kalman filter is devised. The validity of the devised equation is then testified by a simulated example. |
| doi_str_mv | 10.1109/72.737502 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_pubmed_primary_18252512</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>737502</ieee_id><sourcerecordid>27038429</sourcerecordid><originalsourceid>FETCH-LOGICAL-c424t-45b9b8b5361bf8048a6d74743d32d5324aae0c91a4c7b020e9f6c947afe579eb3</originalsourceid><addsrcrecordid>eNqF0T1LxEAQBuBFFL8LWwtJIYpFdHZ2N5stRc4PPLhG67BJJhpNNudugvjvzXFBO61mYB7eKV7Gjjhccg7mSuOlFloBbrBdbiSPAYzYHHeQKjaIeofthfAGwKWCZJvt8BQVKo67bLZwUf9K0aNtWuuiqm568rV7iVrqX7syql3kaPC2GUf_2fn3qPe2dithXRkt_bDaD9hWZZtAh9PcZ8-3s6eb-3i-uHu4uZ7HhUTZx1LlJk9zJRKeVynI1CalllqKUmCpBEprCQrDrSx0DghkqqQwUtuKlDaUi312vs5d-u5joNBnbR0KahrrqBtCZrgxnGsQ_0otBGKqk5U8-1NiKhMJgP_D8W8q0YzwYg0L34XgqcqWvm6t_8o4ZKvCMo3ZurDRnkyhQ95S-SunhkZwOgEbCttU3rqiDr8ukXyMGtnxmtVE9HOdnnwDxeyi4g</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>27038429</pqid></control><display><type>article</type><title>On the Kalman filtering method in neural network training and pruning</title><source>IEEE Electronic Library (IEL)</source><creator>Sum, J. ; Chi-Sing Leung ; Young, G.H. ; Wing-Kay Kan</creator><creatorcontrib>Sum, J. ; Chi-Sing Leung ; Young, G.H. ; Wing-Kay Kan</creatorcontrib><description>In the use of the extended Kalman filter approach in training and pruning a feedforward neural network, one usually encounters the problems of how to set the initial condition and how to use the result obtained to prune a neural network. In this paper, some cues on the setting of the initial condition are presented with a simple example illustrated. Then based on three assumptions: 1) the size of training set is large enough; 2) the training is able to converge; and 3) the trained network model is close to the actual one, an elegant equation linking the error sensitivity measure (the saliency) and the result obtained via an extended Kalman filter is devised. The validity of the devised equation is then testified by a simulated example.</description><identifier>ISSN: 1045-9227</identifier><identifier>EISSN: 1941-0093</identifier><identifier>DOI: 10.1109/72.737502</identifier><identifier>PMID: 18252512</identifier><identifier>CODEN: ITNNEP</identifier><language>eng</language><publisher>New York, NY: IEEE</publisher><subject>Applied sciences ; Biological neural networks ; Computer science ; Computer simulation ; Covariance matrix ; Electric, optical and optoelectronic circuits ; Electronics ; Equations ; Exact sciences and technology ; Extended Kalman filter ; Feedforward neural networks ; Filtering ; Initial conditions ; Kalman filters ; Mathematical analysis ; Mathematical models ; Multilayer perceptrons ; Neural networks ; Pruning ; Testing ; Training</subject><ispartof>IEEE transactions on neural networks, 1999-01, Vol.10 (1), p.161-166</ispartof><rights>1999 INIST-CNRS</rights><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c424t-45b9b8b5361bf8048a6d74743d32d5324aae0c91a4c7b020e9f6c947afe579eb3</citedby><cites>FETCH-LOGICAL-c424t-45b9b8b5361bf8048a6d74743d32d5324aae0c91a4c7b020e9f6c947afe579eb3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/737502$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,776,780,792,4010,27900,27901,27902,54733</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/737502$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=1641737$$DView record in Pascal Francis$$Hfree_for_read</backlink><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/18252512$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Sum, J.</creatorcontrib><creatorcontrib>Chi-Sing Leung</creatorcontrib><creatorcontrib>Young, G.H.</creatorcontrib><creatorcontrib>Wing-Kay Kan</creatorcontrib><title>On the Kalman filtering method in neural network training and pruning</title><title>IEEE transactions on neural networks</title><addtitle>TNN</addtitle><addtitle>IEEE Trans Neural Netw</addtitle><description>In the use of the extended Kalman filter approach in training and pruning a feedforward neural network, one usually encounters the problems of how to set the initial condition and how to use the result obtained to prune a neural network. In this paper, some cues on the setting of the initial condition are presented with a simple example illustrated. Then based on three assumptions: 1) the size of training set is large enough; 2) the training is able to converge; and 3) the trained network model is close to the actual one, an elegant equation linking the error sensitivity measure (the saliency) and the result obtained via an extended Kalman filter is devised. The validity of the devised equation is then testified by a simulated example.</description><subject>Applied sciences</subject><subject>Biological neural networks</subject><subject>Computer science</subject><subject>Computer simulation</subject><subject>Covariance matrix</subject><subject>Electric, optical and optoelectronic circuits</subject><subject>Electronics</subject><subject>Equations</subject><subject>Exact sciences and technology</subject><subject>Extended Kalman filter</subject><subject>Feedforward neural networks</subject><subject>Filtering</subject><subject>Initial conditions</subject><subject>Kalman filters</subject><subject>Mathematical analysis</subject><subject>Mathematical models</subject><subject>Multilayer perceptrons</subject><subject>Neural networks</subject><subject>Pruning</subject><subject>Testing</subject><subject>Training</subject><issn>1045-9227</issn><issn>1941-0093</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>1999</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNqF0T1LxEAQBuBFFL8LWwtJIYpFdHZ2N5stRc4PPLhG67BJJhpNNudugvjvzXFBO61mYB7eKV7Gjjhccg7mSuOlFloBbrBdbiSPAYzYHHeQKjaIeofthfAGwKWCZJvt8BQVKo67bLZwUf9K0aNtWuuiqm568rV7iVrqX7syql3kaPC2GUf_2fn3qPe2dithXRkt_bDaD9hWZZtAh9PcZ8-3s6eb-3i-uHu4uZ7HhUTZx1LlJk9zJRKeVynI1CalllqKUmCpBEprCQrDrSx0DghkqqQwUtuKlDaUi312vs5d-u5joNBnbR0KahrrqBtCZrgxnGsQ_0otBGKqk5U8-1NiKhMJgP_D8W8q0YzwYg0L34XgqcqWvm6t_8o4ZKvCMo3ZurDRnkyhQ95S-SunhkZwOgEbCttU3rqiDr8ukXyMGtnxmtVE9HOdnnwDxeyi4g</recordid><startdate>199901</startdate><enddate>199901</enddate><creator>Sum, J.</creator><creator>Chi-Sing Leung</creator><creator>Young, G.H.</creator><creator>Wing-Kay Kan</creator><general>IEEE</general><general>Institute of Electrical and Electronics Engineers</general><scope>RIA</scope><scope>RIE</scope><scope>IQODW</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><scope>7X8</scope><scope>7SP</scope><scope>F28</scope><scope>FR3</scope></search><sort><creationdate>199901</creationdate><title>On the Kalman filtering method in neural network training and pruning</title><author>Sum, J. ; Chi-Sing Leung ; Young, G.H. ; Wing-Kay Kan</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c424t-45b9b8b5361bf8048a6d74743d32d5324aae0c91a4c7b020e9f6c947afe579eb3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>1999</creationdate><topic>Applied sciences</topic><topic>Biological neural networks</topic><topic>Computer science</topic><topic>Computer simulation</topic><topic>Covariance matrix</topic><topic>Electric, optical and optoelectronic circuits</topic><topic>Electronics</topic><topic>Equations</topic><topic>Exact sciences and technology</topic><topic>Extended Kalman filter</topic><topic>Feedforward neural networks</topic><topic>Filtering</topic><topic>Initial conditions</topic><topic>Kalman filters</topic><topic>Mathematical analysis</topic><topic>Mathematical models</topic><topic>Multilayer perceptrons</topic><topic>Neural networks</topic><topic>Pruning</topic><topic>Testing</topic><topic>Training</topic><toplevel>online_resources</toplevel><creatorcontrib>Sum, J.</creatorcontrib><creatorcontrib>Chi-Sing Leung</creatorcontrib><creatorcontrib>Young, G.H.</creatorcontrib><creatorcontrib>Wing-Kay Kan</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>Pascal-Francis</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><collection>MEDLINE - Academic</collection><collection>Electronics & Communications Abstracts</collection><collection>ANTE: Abstracts in New Technology & Engineering</collection><collection>Engineering Research Database</collection><jtitle>IEEE transactions on neural networks</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Sum, J.</au><au>Chi-Sing Leung</au><au>Young, G.H.</au><au>Wing-Kay Kan</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>On the Kalman filtering method in neural network training and pruning</atitle><jtitle>IEEE transactions on neural networks</jtitle><stitle>TNN</stitle><addtitle>IEEE Trans Neural Netw</addtitle><date>1999-01</date><risdate>1999</risdate><volume>10</volume><issue>1</issue><spage>161</spage><epage>166</epage><pages>161-166</pages><issn>1045-9227</issn><eissn>1941-0093</eissn><coden>ITNNEP</coden><abstract>In the use of the extended Kalman filter approach in training and pruning a feedforward neural network, one usually encounters the problems of how to set the initial condition and how to use the result obtained to prune a neural network. In this paper, some cues on the setting of the initial condition are presented with a simple example illustrated. Then based on three assumptions: 1) the size of training set is large enough; 2) the training is able to converge; and 3) the trained network model is close to the actual one, an elegant equation linking the error sensitivity measure (the saliency) and the result obtained via an extended Kalman filter is devised. The validity of the devised equation is then testified by a simulated example.</abstract><cop>New York, NY</cop><pub>IEEE</pub><pmid>18252512</pmid><doi>10.1109/72.737502</doi><tpages>6</tpages></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | ISSN: 1045-9227 |
| ispartof | IEEE transactions on neural networks, 1999-01, Vol.10 (1), p.161-166 |
| issn | 1045-9227 1941-0093 |
| language | eng |
| recordid | cdi_pubmed_primary_18252512 |
| source | IEEE Electronic Library (IEL) |
| subjects | Applied sciences Biological neural networks Computer science Computer simulation Covariance matrix Electric, optical and optoelectronic circuits Electronics Equations Exact sciences and technology Extended Kalman filter Feedforward neural networks Filtering Initial conditions Kalman filters Mathematical analysis Mathematical models Multilayer perceptrons Neural networks Pruning Testing Training |
| title | On the Kalman filtering method in neural network training and pruning |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-01T21%3A14%3A05IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=On%20the%20Kalman%20filtering%20method%20in%20neural%20network%20training%20and%20pruning&rft.jtitle=IEEE%20transactions%20on%20neural%20networks&rft.au=Sum,%20J.&rft.date=1999-01&rft.volume=10&rft.issue=1&rft.spage=161&rft.epage=166&rft.pages=161-166&rft.issn=1045-9227&rft.eissn=1941-0093&rft.coden=ITNNEP&rft_id=info:doi/10.1109/72.737502&rft_dat=%3Cproquest_RIE%3E27038429%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=27038429&rft_id=info:pmid/18252512&rft_ieee_id=737502&rfr_iscdi=true |