Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Machine Learning Methods for Evaluating Public Crisis: Meta-Analysis

Izunna Okpala, Shane Halse, Jess Kropczynski

TL;DR

This paper addresses how machine learning is used to evaluate public crises by synthesizing existing studies. It employs a meta-review of 55 peer-reviewed articles from IEEE Xplore, ScienceDirect, and ISCRAM published between 2010 and 2021 to identify prevalent ML tasks, data sources, and algorithms used in crisis management. The analysis finds a strong emphasis on supervised learning and classification, with SVM, neural networks, Naive Bayes, and Random Forest as prominent algorithms, and social media data as a common data source. The results highlight trends such as rising interest in crisis informatics and health-related crises (notably COVID-19) and increasing use of NLP and sentiment analysis, offering guidance for researchers and practitioners on where ML methods are most effectively applied in crisis contexts.

Abstract

This study examines machine learning methods used in crisis management. Analyzing detected patterns from a crisis involves the collection and evaluation of historical or near-real-time datasets through automated means. This paper utilized the meta-review method to analyze scientific literature that utilized machine learning techniques to evaluate human actions during crises. Selected studies were condensed into themes and emerging trends using a systematic literature evaluation of published works accessed from three scholarly databases. Results show that data from social media was prominent in the evaluated articles with 27% usage, followed by disaster management, health (COVID) and crisis informatics, amongst many other themes. Additionally, the supervised machine learning method, with an application of 69% across the board, was predominant. The classification technique stood out among other machine learning tasks with 41% usage. The algorithms that played major roles were the Support Vector Machine, Neural Networks, Naive Bayes, and Random Forest, with 23%, 16%, 15%, and 12% contributions, respectively.

Machine Learning Methods for Evaluating Public Crisis: Meta-Analysis

TL;DR

This paper addresses how machine learning is used to evaluate public crises by synthesizing existing studies. It employs a meta-review of 55 peer-reviewed articles from IEEE Xplore, ScienceDirect, and ISCRAM published between 2010 and 2021 to identify prevalent ML tasks, data sources, and algorithms used in crisis management. The analysis finds a strong emphasis on supervised learning and classification, with SVM, neural networks, Naive Bayes, and Random Forest as prominent algorithms, and social media data as a common data source. The results highlight trends such as rising interest in crisis informatics and health-related crises (notably COVID-19) and increasing use of NLP and sentiment analysis, offering guidance for researchers and practitioners on where ML methods are most effectively applied in crisis contexts.

Abstract

This study examines machine learning methods used in crisis management. Analyzing detected patterns from a crisis involves the collection and evaluation of historical or near-real-time datasets through automated means. This paper utilized the meta-review method to analyze scientific literature that utilized machine learning techniques to evaluate human actions during crises. Selected studies were condensed into themes and emerging trends using a systematic literature evaluation of published works accessed from three scholarly databases. Results show that data from social media was prominent in the evaluated articles with 27% usage, followed by disaster management, health (COVID) and crisis informatics, amongst many other themes. Additionally, the supervised machine learning method, with an application of 69% across the board, was predominant. The classification technique stood out among other machine learning tasks with 41% usage. The algorithms that played major roles were the Support Vector Machine, Neural Networks, Naive Bayes, and Random Forest, with 23%, 16%, 15%, and 12% contributions, respectively.
Paper Structure (12 sections, 6 figures, 3 tables)

This paper contains 12 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Research Questions
  3. Background
  4. Machine Learning for Crisis Detection and Management
  5. Methodology
  6. Search technique
  7. Data Selection
  8. Data Extraction
  9. Result
  10. Discussion
  11. Limitations
  12. Conclusion

Figures (6)

  • Figure 1: Data search and selection Algorithm
  • Figure 2: Inclusion and Exclusion flow diagram
  • Figure 3: Graph representation of the ML methods and tasks
  • Figure 4: Distribution of keywords across reviewed articles
  • Figure 5: Algorithms applied in reviewed articles (heatmap)
  • ...and 1 more figures