[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-240817-105":3,"doc-detail-240817-en":81,"detail-sidebar-cat-1-en-105":98},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","classification-of-bioacoustic-time-series-training-a-decision-template-fusion-mapping","Classification of Bioacoustic Time Series - Training a Decision Template Fusion Mapping","","Adaptable fusion techniques for combining local classifier decisions computed from different feature subspaces drive this paper. Decision template fusion is discussed within neural network learning algorithms and then applied to recognition of bioacoustic time series. A two-layer fusion architecture is presented, contrasting static fusion mapping with adaptable mappings that learn classifier outputs and temporal integration for final class decisions.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/template/","Template",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/template/presentations/","Presentations",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/template/classification-of-bioacoustic-time-series-training-a-decision-template-fusion-mapping/240817/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/classification-of-bioacoustic-time-series-training-a-decision-template-fusion-mapping/240817.png","ImageObject",442,249,{"name":42,"@type":43},"Angel","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-11-01","2026-09-11",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",13,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What problem does decision template fusion address in this paper?","Question",{"text":63,"@type":64},"It focuses on combining local classifier decisions derived from different feature subspaces for improved time-series classification.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"How does the CDT fusion architecture compute the final classification?",{"text":68,"@type":64},"It performs per-feature classification (C-step), fuses local decisions across time windows (D-step), and integrates decisions temporally (T-step) to select the class with maximum membership.",{"name":70,"@type":61,"acceptedAnswer":71},"What role do decision templates play in the method?",{"text":72,"@type":64},"Decision templates act as a trainable aggregation rule, serving as characteristic outputs for each class and enabling similarity-based membership computation.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},240817,1791492111,{"code":4,"msg":82,"data":83},"success",{"doc_id":79,"user_id":84,"nickname":42,"user_avatar":85,"doc_module":22,"category_id":86,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":87,"file_id":88,"file_url":89,"file_type":90,"file_size":91,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":92,"language":93,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":94,"faqs":95,"seo_title":96,"seo_description":12,"update_tm":97,"read_time":26},687207412472,"https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",11,"ESANN'2004 proceedings-European Symposium on Artificial Neural Networks Bruges (Belgium), 28-30 April 2004, d-side publi., ISBN 2-930307-04-8, pp. 555-560  \nClassi􀀌cation of Bioacoustic Time Series by  \nTraining a Decision Template Fusion Mapping  \nF. Schwenker, C. Dietrich, G. Palm  \nUniversity of Ulm, D-89069 Ulm, Germany  \nAbstract. Adaptable fusion techniques for the combination of local classi􀀌er decisions calculated from di􀀋erent feature subspaces are the topic of this paper. Decision template fusion is is discussed in the context of neural network learning algorithms, and applied to the recognition of bioacoustic time series.  \n1 Introduction  \nCombining the classi􀀌cation powers of several classi􀀌ers is regarded as a general problem in various pattern recognition applications [9, 4] . For the decison fusion static and adaptable combining paradigms [1, 5] have been proposed and discussed. In the static fusion mapping approach the individual classi􀀌ers of the ensemble are separetely trained by a single pass, and classi􀀌er fusion mapping is implemented by a prede􀀌ned so-called aggregation rule. Adaptable fusion mappings are trained by a two-phase training procedure:  \n1. building the Classi􀀌er Layer consisting of a set of 􀀌rst level classi􀀌ers using training data R, and  \n2. training the Fusion Layer performing a mapping of the classi􀀌er outputs (soft or crisp decisions) into the set of desired class labels by using a validation set V.  \nThe overall classi􀀌er architecture (see Figure 1) is a two-layered structure similar to multilayer perceptrons and radial basis functions.  \n2 Classi􀀌cation Fusion of Local Features  \nIn this Section we propose a static and a trainable fusion architecture for time series classi􀀌cation. It is assumed that each time series is labeled with its corresponding class label ! 2 􀀊, 􀀊 = f1;:::;Lg. A sliding window Wj covering a small part of the time series s (t) is moved over the whole time series. For each window Wj , j = 1 ;:::; J a set of I features xi (j) 2 IRDi , i = 1;:::;I and Di 2 IN, is extracted. Typically J , the number of time windows varies from  \nESANN'2004 proceedings-European Symposium on Artificial Neural Networks Bruges (Belgium), 28-30 April 2004, d-side publi., ISBN 2-930307-04-8, pp. 555-560  \nFeature 1  \nFeature i  \nFeature I  \nClassifier Layer Fusion Layer  \nClassification Transformation  \nFigure 1: MCS with classi􀀌er layer and fusion layer. The combination of the classi􀀌er outputs Ci (xi ), i = 1;:::;I is accomplished through a fusion mapping F (C 1 (x1 ) ;:::; CI (xI )) . We consider fusion mappings where the classi􀀌er output Ci (xi ) are linearly combined.  \ntime series to time series. For s (t) this leads to I feature streams xi (j)Jj=1 . To determine the class membership of an input vector x to all L classes, a fuzzyk-nearest-neighbour classi􀀌er is used [8] . Such a fuzzy classi􀀌er C is de􀀌ned as a mapping C : IRD ! [0 ; 1]L , i.e. , the output C (x) = (C1 (x);:::; CL (x)) contains the memberships of x to each class. After normalization by Cl (x) :=􀀎l (x)= P 􀀎l (x) we get soft labels C (x) 2 􀀁 with  \nL  \n􀀁 := f(y 1 ;:::; yL ) 2 [0; 1]L j X yl = 1g (1)  \nl=1  \n2.1 CDT-Architecture  \nThe CDT fusion performs the classi􀀌cation in three steps:  \n1.) Classi􀀌cation of single feature vectors (C-step)  \nFor each feature i = 1;:::;I a classi􀀌er Ci is given through a mapping  \nCi : IRDi ! 􀀁 : (2) Thus, for each time window Wj , j = 1 ;:::; J , I classi􀀌cation results C 1 (x1 (j));:::; CI (xI (j)) based on the individual features x 1 (j);:::; xI (j) are calculated.  \n2.) Decision fusion of the local decisions (D-step)  \nFor each time window Wj the I classi􀀌cation results are combined into  \nESANN'2004 proceedings-European Symposium on Artificial Neural Networks Bruges (Belgium), 28-30 April 2004, d-side publi., ISBN 2-930307-04-8, pp. 555-560  \na local decision zj 2 􀀁 through a fusion mapping F : 􀀁 I ! 􀀁  \nzj := F (C 1 (x1 (j));:::; CI (xI (j))); j = 1 ;:::; J: (3)  \n3.) Temporal fu","cbCaisKNOWIkqqHb","https://ap.wps.com/l/cbCaisKNOWIkqqHb","pdf",245064,6,"English","# Introduction\n## Classification Fusion of Local Features\n## CDT-Architecture\n## Decision Templates","[{\"question\":\"What problem does decision template fusion address in this paper?\",\"answer\":\"It focuses on combining local classifier decisions derived from different feature subspaces for improved time-series classification.\"},{\"question\":\"How does the CDT fusion architecture compute the final classification?\",\"answer\":\"It performs per-feature classification (C-step), fuses local decisions across time windows (D-step), and integrates decisions temporally (T-step) to select the class with maximum membership.\"},{\"question\":\"What role do decision templates play in the method?\",\"answer\":\"Decision templates act as a trainable aggregation rule, serving as characteristic outputs for each class and enabling similarity-based membership computation.\"}]","Classification of Bioacoustic Time Series - Training a Decision Template Fusion Mapping | PDF",1789166140,{"code":4,"msg":82,"data":99},[100,103,108,113,118,123,128,133,138],{"id":86,"doc_module":22,"doc_module_name":25,"category_name":29,"show_sort_weight":101,"slug":102},90,"presentations",{"id":104,"doc_module":22,"doc_module_name":25,"category_name":105,"show_sort_weight":106,"slug":107},12,"Resumes",80,"resumes",{"id":109,"doc_module":22,"doc_module_name":25,"category_name":110,"show_sort_weight":111,"slug":112},14,"Invoices",70,"invoices",{"id":114,"doc_module":22,"doc_module_name":25,"category_name":115,"show_sort_weight":116,"slug":117},15,"Posters",60,"posters",{"id":119,"doc_module":22,"doc_module_name":25,"category_name":120,"show_sort_weight":121,"slug":122},16,"Social Media",50,"social-media",{"id":124,"doc_module":22,"doc_module_name":25,"category_name":125,"show_sort_weight":126,"slug":127},17,"Forms",40,"forms",{"id":129,"doc_module":22,"doc_module_name":25,"category_name":130,"show_sort_weight":131,"slug":132},18,"Letters",30,"letters",{"id":134,"doc_module":22,"doc_module_name":25,"category_name":135,"show_sort_weight":136,"slug":137},21,"Paper Templates",5,"papers-templates",{"id":139,"doc_module":22,"doc_module_name":25,"category_name":140,"show_sort_weight":4,"slug":141},158,"General","general-158"]