Abstract
Constructed Wetlands (CWs) are a nature-based solution for the treatment of wastewater. The CWetlands - the Constructed Wetlands Knowledge Platform - intends to help understand how CWs can support achieving the Sustainable Development Goals. The platform is based on more than 100 attributes of CWs including criteria for design criteria, operation, efficiency, climate and other geographical factors. This study aims at developing an attribute value extraction mechanism tool in R to extract meaningful information from peer-reviewed journal articles in a reliable and fast way.
The tool focuses on the extraction of eighteen different extractable attributes gathered in 4 classes, which describe the main characteristics of CW systems.
The process contains 4 sub-processes: 1-2) the papers are accessed and pre-processed, 3) the attributes are extracted by two data mining techniques: Keyword Match and Web Scrap, and 4) the values are exported to a database.
For the development and testing of the tool,13 articles were used. The tool achieved a mean success rate of 79% in 30 min; less compared with the 480 min needed with a manual approach. In further versions, the tool is expected to obtain a higher success rate in all attributes. (C) 2019 The Author(s). Published by Elsevier B.V.