Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methods
Whether triggered by natural or human-made events, wildfires are considered one of the most traumatic events to our community and environment. Thus, properly predicting wildfires continues to be an active area of research. This work showcases a statistical overview of the problem of wildfires and th...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
KeAi Communications Co. Ltd.
2022-09-01
|
| Series: | Natural Hazards Research |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666592122000373 |
| _version_ | 1811199555936452608 |
|---|---|
| author | Mohammad Khaled Al-Bashiti M.Z. Naser |
| author_facet | Mohammad Khaled Al-Bashiti M.Z. Naser |
| author_sort | Mohammad Khaled Al-Bashiti |
| collection | DOAJ |
| description | Whether triggered by natural or human-made events, wildfires are considered one of the most traumatic events to our community and environment. Thus, properly predicting wildfires continues to be an active area of research. This work showcases a statistical overview of the problem of wildfires and then presents a dense data-driven (D3) approach that leverages a variety of machine learning (ML) techniques, namely, blackbox and eXplainable ML (i.e., deep learning (DL), decision tree (DT), Stochastic Gradient Descent (SGD), Extreme Gradient Boosted Trees (ExGBT), Logistic regression (LR)), and symbolic ML via genetic algorithms (GA) to classify and predict wildfire breakouts. This approach was developed and validated using two databases comprising more than 1.04 million geo-referenced wildfires that burned over 359,000 km2 (88.7 million acres) between 1992 and 2015 in North America and Europe. Despite the complex nature of wildfire formation and the interdependency of its governing factors, the findings of this D3 analysis show the feasibility of utilizing ML in preciously classifying the expected size of wildfires and predicting the possibility of the breakout of wildfires. |
| first_indexed | 2024-04-12T01:50:39Z |
| format | Article |
| id | doaj.art-9a8df14d98af45fdad2a4066cc6b60fa |
| institution | Directory Open Access Journal |
| issn | 2666-5921 |
| language | English |
| last_indexed | 2024-04-12T01:50:39Z |
| publishDate | 2022-09-01 |
| publisher | KeAi Communications Co. Ltd. |
| record_format | Article |
| series | Natural Hazards Research |
| spelling | doaj.art-9a8df14d98af45fdad2a4066cc6b60fa2022-12-22T03:52:57ZengKeAi Communications Co. Ltd.Natural Hazards Research2666-59212022-09-0123154165Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methodsMohammad Khaled Al-Bashiti0M.Z. Naser1School of Civil & Environmental Engineering and Earth Sciences (SCEEES), Clemson University, USASchool of Civil & Environmental Engineering and Earth Sciences (SCEEES), Clemson University, USA; AI Research Institute for Science and Engineering (AIRISE), Clemson University, Clemson, SC, 29634, USA; Corresponding author. School of Civil & Environmental Engineering and Earth Sciences (SCEEES), Clemson University, USA.Whether triggered by natural or human-made events, wildfires are considered one of the most traumatic events to our community and environment. Thus, properly predicting wildfires continues to be an active area of research. This work showcases a statistical overview of the problem of wildfires and then presents a dense data-driven (D3) approach that leverages a variety of machine learning (ML) techniques, namely, blackbox and eXplainable ML (i.e., deep learning (DL), decision tree (DT), Stochastic Gradient Descent (SGD), Extreme Gradient Boosted Trees (ExGBT), Logistic regression (LR)), and symbolic ML via genetic algorithms (GA) to classify and predict wildfire breakouts. This approach was developed and validated using two databases comprising more than 1.04 million geo-referenced wildfires that burned over 359,000 km2 (88.7 million acres) between 1992 and 2015 in North America and Europe. Despite the complex nature of wildfire formation and the interdependency of its governing factors, the findings of this D3 analysis show the feasibility of utilizing ML in preciously classifying the expected size of wildfires and predicting the possibility of the breakout of wildfires.http://www.sciencedirect.com/science/article/pii/S2666592122000373WildfiresForestsMachine learningBig dataExplainable MLSymbolic ML |
| spellingShingle | Mohammad Khaled Al-Bashiti M.Z. Naser Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methods Natural Hazards Research Wildfires Forests Machine learning Big data Explainable ML Symbolic ML |
| title | Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methods |
| title_full | Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methods |
| title_fullStr | Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methods |
| title_full_unstemmed | Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methods |
| title_short | Machine learning for wildfire classification: Exploring blackbox, eXplainable, symbolic, and SMOTE methods |
| title_sort | machine learning for wildfire classification exploring blackbox explainable symbolic and smote methods |
| topic | Wildfires Forests Machine learning Big data Explainable ML Symbolic ML |
| url | http://www.sciencedirect.com/science/article/pii/S2666592122000373 |
| work_keys_str_mv | AT mohammadkhaledalbashiti machinelearningforwildfireclassificationexploringblackboxexplainablesymbolicandsmotemethods AT mznaser machinelearningforwildfireclassificationexploringblackboxexplainablesymbolicandsmotemethods |