Abstract:
Landslides are among the natural disasters that most affect the population in Brazil and in the world, as they are responsible for numerous human, economic, social and environmental losses. Highresolution aerial and satellite imagery, that contribute to identify this type of event are valuable sources of information, even if they are not the only ones. Machine learning (ML) techniques are often adopted to identify relevant patterns, as high performance computing and data visualization are widely and successfully applied to natural disaster-related data, particularly for landslides. The greater the volume of data processed during machine learning, the better the accuracy of the results obtained. Among other reasons, Big Data (BD) has stood out for its ability to integrate and generate a large volume of data from different sources and in different formats. The use of BD concepts and techniques has become common in several areas of knowledge, in particular for the storage and analysis of data related to natural accidents and, thus, helping to improve decision making as well as prevention of disasters. With the emergence of the BD, it came the challenge of aligning decisionmaking processes, so that there is no shortage of information or useless data for decision makers. In the case of Brazil and considering, especially, applications to landslide data, there is a lack of articles dealing with the subject. The aim of this study is to investigate the main features used on machine learning techniques, to map the literature, thus pointing out new routes and opportunities in the field. In the systematic review, the eligibility criteria adopted are studies published in the last 4 years (from January 2015 to December 2019), in English, at national and international levels, including publications in scientific events and relevant journals, through databases such as Scopus and Web of Science. The search was performed in a structured form with the terms "Big Data*", "machine learn*","landslid*", "forecast*" and "predict*" and the like through the boolean operator “AND” or “OR” in order to restrict the theme to the aspects to be discussed. The researches were included when they contained a description of the BD and ML techniques, and could be applied to landslide studies.
Referência:
AZEVEDO, Caio da Silva; CORREIA, Thais Passos; CORSI, Alessandra Cristina; SPINOLA, Mauro de Mesquita; QUINTANILHA, José Alberto. A systematic review: big data analytic and landslides. In: INTERNATIONAL SYMPOSIUM ON LANDSLIDES, 13., 2021, Colombia. On-line. Proceedings on-line… 9p..
Access to the work presented at the event, on the website:
https://www.issmge.org/uploads/publications/105/106/ISL2020-31.pdf
Landslides are among the natural disasters that most affect the population in Brazil and in the world, as they are responsible for numerous human, economic, social and environmental losses. Highresolution aerial and satellite imagery, that contribute to identify this type of event are valuable sources of information, even if they are not the only ones. Machine learning (ML) techniques are often adopted to identify relevant patterns, as high performance computing and data visualization are widely and successfully applied to natural disaster-related data, particularly for landslides. The greater the volume of data processed during machine learning, the better the accuracy of the results obtained. Among other reasons, Big Data (BD) has stood out for its ability to integrate and generate a large volume of data from different sources and in different formats. The use of BD concepts and techniques has become common in several areas of knowledge, in particular for the storage and analysis of data related to natural accidents and, thus, helping to improve decision making as well as prevention of disasters. With the emergence of the BD, it came the challenge of aligning decisionmaking processes, so that there is no shortage of information or useless data for decision makers. In the case of Brazil and considering, especially, applications to landslide data, there is a lack of articles dealing with the subject. The aim of this study is to investigate the main features used on machine learning techniques, to map the literature, thus pointing out new routes and opportunities in the field. In the systematic review, the eligibility criteria adopted are studies published in the last 4 years (from January 2015 to December 2019), in English, at national and international levels, including publications in scientific events and relevant journals, through databases such as Scopus and Web of Science. The search was performed in a structured form with the terms "Big Data*", "machine learn*","landslid*", "forecast*" and "predict*" and the like through the boolean operator “AND” or “OR” in order to restrict the theme to the aspects to be discussed. The researches were included when they contained a description of the BD and ML techniques, and could be applied to landslide studies.
Referência:
AZEVEDO, Caio da Silva; CORREIA, Thais Passos; CORSI, Alessandra Cristina; SPINOLA, Mauro de Mesquita; QUINTANILHA, José Alberto. A systematic review: big data analytic and landslides. In: INTERNATIONAL SYMPOSIUM ON LANDSLIDES, 13., 2021, Colombia. On-line. Proceedings on-line… 9p..
Access to the work presented at the event, on the website:
https://www.issmge.org/uploads/publications/105/106/ISL2020-31.pdf
