Return to search

Estudo de comunidades de dípteros necrófagos sob o formalismo de grafos e árvores de decisão /

Título original da capa: Estudo de comunidades de dípteros necrófagos sob o formalismo de biclusters e árvores de decisão / Orientador: Claudio José Von Zuben / Banca: Patricia Jacqueline Thyssen / Banca: Edilberto Giannotti / Banca: Maria José de Oliveira Campos / Banca: Wesley Augusto Conde Godoy / Resumo: The pattern recognition in ecological communities is one of the most antique and persistent challenges of ecological science. In this sense, through mathematical modelling from graphs, using biclusters and decision trees, the objective of the present work was to find patterns in the occurrence of necrophagous dipterans collected in three different areas (rural, urban and forest) and with three different baits (sardines, liver and beef) in the city of Rio Claro-SP. Algorithms (InClose and RInClose-biclusters, C 4.5-decision trees) were used to obtain the results. Through the analysis of biclusters, it can be highlighted those that related Atherigona orientalis (Schiner) and Musca domestica Linnaeus in all environments, less forest area, probably due to the fact of being species known for their preference for areas inhabited by humans. On the other hand, even being species possessing high dependence by antropobiocenose, other biclusters indicated an association between Lucilia eximia (Wiedemann) and Chrysomya megacephala (Fabricius) occurring with greater frequency in the forest, attracted by the bait of meat or liver, both known for their importance in the attractiveness of calliphorid fly. This may be explained due to the population growth occurred around the forest, giving the characteristics of place inhabited by man. With regard to decision trees, were obtained in total 19 decision trees (8 to the family Sarcophagidae, 4 for Muscidade and 7 to Calliphoridae) and from them... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The pattern recognition in ecological communities is one of the most antique and persistent challenges of ecological science. In this sense, through mathematical modelling from graphs, using biclusters and decision trees, the objective of the present work was to find patterns in the occurrence of necrophagous dipterans collected in three different areas (rural, urban and forest) and with three different baits (sardines, liver and beef) in the city of Rio Claro-SP. Algorithms (InClose and RInClose-biclusters, C 4.5-decision trees) were used to obtain the results. Through the analysis of biclusters, it can be highlighted those that related Atherigona orientalis (Schiner) and Musca domestica Linnaeus in all environments, less forest area, probably due to the fact of being species known for their preference for areas inhabited by humans. On the other hand, even being species possessing high dependence by antropobiocenose, other biclusters indicated an association between Lucilia eximia (Wiedemann) and Chrysomya megacephala (Fabricius) occurring with greater frequency in the forest, attracted by the bait of meat or liver, both known for their importance in the attractiveness of calliphorid fly. This may be explained due to the population growth occurred around the forest, giving the characteristics of place inhabited by man. With regard to decision trees, were obtained in total 19 decision trees (8 to the family Sarcophagidae, 4 for Muscidade and 7 to Calliphoridae) and from them various patterns can be detected such as: A. orientalis, Synthesiomyia nudiseta (Wulp, 1883) and M. domestica being present together if the season is Spring, the bait is sardines and the location is urban. Oxysarcodexia thornax (Walker) was the only specie in which to achieve the condition of presence, it is necessary to walk only one node ... (Complete abstract electronic access below) / Doutor

Identiferoai:union.ndltd.org:UNESP/oai:www.athena.biblioteca.unesp.br:UEP01-000871814
Date January 2016
CreatorsSouza, Caroline Rodrigues de.
ContributorsUniversidade Estadual Paulista "Júlio de Mesquita Filho" Instituto de Biociências (Campus de Rio Claro).
PublisherRio Claro,
Source SetsUniversidade Estadual Paulista
LanguageMultiple languages, English
Detected LanguageEnglish
Typetext
Format181 f. :
Coverages-bl-sp
RelationSistema requerido: Adobe Acrobat Reader

Page generated in 0.0017 seconds