Author(s): Leticia da Costa e Silva | Denise Fukumi Tsunoda | Viviane Deslandes
Journal: AtoZ : Novas Práticas em Informação e Conhecimento
ISSN 2237-826X
Volume: 1;
Issue: 1;
Start page: 61;
Date: 2011;
Original page
Keywords: Mineração de dados | Bases de dados | Aves florestais | Thamnophilidae (ave) | Cantos dos pássaros
ABSTRACT
Introduction: Describes the use of a data mining technique about the song, biology and micro-habitat of the Thamnophilidae bird family in order to find patterns which relate them.Method: A database was built in Excel ® spreadsheet listing 82 species of the family of the bird Thamnophilidae comprising various attributes related to bird calling features, biology and micro-habitat in which they are found. For the analysis it was used the algorithm APRIORI in the WEKA 3.7.1 software.Results: The association of the different attributes of the 82 different species, considering 10% of minimum support and 90% of minimum confidence, allowed the rescued of 172 patterns, from which 42 contained one of the song’s attributes: PC1 e PC2. The patterns which related the attribute PC2 were the most expressive ones due to its relation to the size and gender of the family.Conclusions: The experiment demonstrated that the algorithm could be better suited in larger databases and/or when the data standardization presents a lower number of categories, what could be a limitation in the macroecology field. Nonetheless, it has presented itself as an alternative instrument to the exploratory study of the relations among diverse attributes, which results could serve as objects for further analysis.
Journal: AtoZ : Novas Práticas em Informação e Conhecimento
ISSN 2237-826X
Volume: 1;
Issue: 1;
Start page: 61;
Date: 2011;
Original page
Keywords: Mineração de dados | Bases de dados | Aves florestais | Thamnophilidae (ave) | Cantos dos pássaros
ABSTRACT
Introduction: Describes the use of a data mining technique about the song, biology and micro-habitat of the Thamnophilidae bird family in order to find patterns which relate them.Method: A database was built in Excel ® spreadsheet listing 82 species of the family of the bird Thamnophilidae comprising various attributes related to bird calling features, biology and micro-habitat in which they are found. For the analysis it was used the algorithm APRIORI in the WEKA 3.7.1 software.Results: The association of the different attributes of the 82 different species, considering 10% of minimum support and 90% of minimum confidence, allowed the rescued of 172 patterns, from which 42 contained one of the song’s attributes: PC1 e PC2. The patterns which related the attribute PC2 were the most expressive ones due to its relation to the size and gender of the family.Conclusions: The experiment demonstrated that the algorithm could be better suited in larger databases and/or when the data standardization presents a lower number of categories, what could be a limitation in the macroecology field. Nonetheless, it has presented itself as an alternative instrument to the exploratory study of the relations among diverse attributes, which results could serve as objects for further analysis.
