Literature mining on dipterocarps: towards better informed natural regeneration and reforestation in Luzon, Philippines
2019
Gabud, R.S. | Batista-Navarro, R.T. | Mariano, V.Y. | Mendoza, E.R. | Yap, S.L.
Access to comprehensive information is critical in Natural Resource Management (NRM). While information resources largely shape the decision-making process of conservation policy regulators, significant barriers could present themselves at any point along the NRM information supply pipeline. Similarly, these barriers could appear in the management of the natural regeneration and reforestation of lowland dipterocarp forests in the Philippines, Moreover, the supra-annual mass flowering of dipterocarps occurs in irregular intervals, thus predicting the likelihood of their natural regeneration, to subsequently make plans for reforestation is an extremely complex task. This study aimed to address information gaps by utilizing text mining, which is the automatic process of extracting structured information from documents such as scientific publications, books, and agency reports. The results were based on a preliminary version of the annotated document collection or corpus. The reliability of the corpus was determined by inter-annotator agreement (IAA). Satisfactory IAA was obtained for taxon and geographic location concept types with 90.35% and 93.82% F-scores, respectively. This corpus could provide NRM policy-makers with a searchable database on the distribution and mass flowering of Philippine dipterocarps for dipterocarp species regeneration and reforestation.
显示更多 [+] 显示较少 [-]