Scholarlink

This article presents a study that combines data mining algorithms with procedural analysis of presence/absence matrices, aiming to generate strategic diagnostics for species conservation in the Galápagos Islands. A reproducible pipeline is developed within the framework of the Knowledge Discovery in Databases (KDD) process, encompassing automated record collection through Selenium, taxonomic standardization using data analysis tools (Pandas), and the application of clustering algorithms such as K-Means and DBSCAN to identify emerging biogeographic patterns. Additionally, a consolidated animal dataset in presence/absence matrix format by island was integrated, enabling the quantification of essential metrics on data quality and distribution. Among the main findings, 107,974 processed data points were recorded, corresponding to 10,248 unique species; 97,726 duplicate records and 46,635 species exclusive to a single island were identified. Island coverage revealed marked heterogeneity, with Santa Cruz (84,820 records), San Cristóbal (38,848), and Isabela (42,651) standing out, compared to peripheral islands like Wolf (1,889) and Genovesa (3,928), which show low representation. In terms of conservation, species classified as least concern (LC, 5,630) predominate, followed by those with deficient data (DD, 4,690) and not evaluated (NE, 4,387). Species at risk are also evident: vulnerable (VU, 3,181), endangered (EN, 1,277) and critically endangered (CR, 1,180), along with records of extinct species (EX, 222; RE, 36; EW, 2). These results reaffirm the richness and vulnerability of the archipelago and highlight the importance of differentiated conservation policies by island. Additionally, the clustering results were validated using silhouette index and density threshold tuning, ensuring analytical consistency. The study emphasizes the application of FAIR and CARE principles to strengthen transparency, traceability, and reproducibility in open biodiversity science.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data Mining Algorithms Analysis for Species Conservation in the Galápagos Islands

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Enrique Nicolás Barnuevo Montaño,
  • Gabriela Soledad Montaño Espinosa,
  • Lorena Elizabeth Conde-Zhingre,
  • Milton Ricardo Palacios Morocho

摘要

This article presents a study that combines data mining algorithms with procedural analysis of presence/absence matrices, aiming to generate strategic diagnostics for species conservation in the Galápagos Islands. A reproducible pipeline is developed within the framework of the Knowledge Discovery in Databases (KDD) process, encompassing automated record collection through Selenium, taxonomic standardization using data analysis tools (Pandas), and the application of clustering algorithms such as K-Means and DBSCAN to identify emerging biogeographic patterns. Additionally, a consolidated animal dataset in presence/absence matrix format by island was integrated, enabling the quantification of essential metrics on data quality and distribution. Among the main findings, 107,974 processed data points were recorded, corresponding to 10,248 unique species; 97,726 duplicate records and 46,635 species exclusive to a single island were identified. Island coverage revealed marked heterogeneity, with Santa Cruz (84,820 records), San Cristóbal (38,848), and Isabela (42,651) standing out, compared to peripheral islands like Wolf (1,889) and Genovesa (3,928), which show low representation. In terms of conservation, species classified as least concern (LC, 5,630) predominate, followed by those with deficient data (DD, 4,690) and not evaluated (NE, 4,387). Species at risk are also evident: vulnerable (VU, 3,181), endangered (EN, 1,277) and critically endangered (CR, 1,180), along with records of extinct species (EX, 222; RE, 36; EW, 2). These results reaffirm the richness and vulnerability of the archipelago and highlight the importance of differentiated conservation policies by island. Additionally, the clustering results were validated using silhouette index and density threshold tuning, ensuring analytical consistency. The study emphasizes the application of FAIR and CARE principles to strengthen transparency, traceability, and reproducibility in open biodiversity science.