Skip to content

Utilizing KNN- Classification to predict the likelihood of forest fires to occur in Algeria. Project based on a summative evaluation from UBC's DSCI 100 course

Notifications You must be signed in to change notification settings

MartinYTSo/dsci_100_algerian_forest_fires_project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

85 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Using Data Science to Predict Fires in the Algerian Forest from Weather Characteristics

Introduction

A forest fire is unplanned and uncontrollable and can occur by lightning or human carelessness in forests, grasslands or shrubbery (Government of Canada, 2020). The vast majority of forest fires are human-caused, however dry climate, hot temperatures, lightning, and volcanic eruption can also lead to their occurrence (National Park Service, 2018). The past decade of climate change has only exacerbated the amount of forest fires, leading to more frequent and extreme occurences.

Wildfire agencies use many variables to indicate an imminent wildfire and the evolution of machine learning has provided us the ability to predict future events by analyzing these variables. Thus, we pose the predictive question: do certain variables allow us to determine if a forest fire has or will occur and if so, how accurate will they be?

To support our hypothesis, we used a dataset on Algerian Forest Fires from UCI (Faroudja & Izeboudjen, 2020). The dataset contains a culmination of forest fire observations and data in two regions of Algeria: the Bejaia region and the Sidi Bel-Abbes region. The timeline of this dataset is from June 2012 to September 2012. In this project, we focused on whether certain weather characteristics could predict forest fires in these regions using the K-NN Classification algorithm and later, we evaluated the accuracy of the model.

Repository Contents:

In this repository, you will find:

  • Forest_Fire_Project.ipynb: For the full project with narration.
  • .ipynb_checkpoints: For previous versions of Forest_Fire_Project.ipynb.

The dataset used can be found here from the UCI Machine Learning Repository.

References

Government of Canada. (2020). Forest fires. Retrieved from https://www.nrcan.gc.ca/our-natural-resources/forests-forestry/wildland-fires-insects-disturban/forest-fires/13143.

National Park Service. (2018). Wildfire Causes and Evaluations. Retrieved from https://www.nps.gov/articles/wildfire-causes-and-evaluation.htm#:~:text=Nearly%2085%20percent*%20of%20wildland,and%20intentional%20acts%20of%20arson.&text=Lightning%20is%20one%20of%20the%20two%20natural%20causes%20of%20fires.

Faroudja, A., Izeboudjen, N. (2020). Predicting forest fire in Algeria using data mining techniques: Case study of the decision tree algorithm. International Conference on Advanced Intelligent Systems for Sustainable Development, 363-370. Retrieved from https://link.springer.com/chapter/10.1007/978-3-030-36674-2_37

About

Utilizing KNN- Classification to predict the likelihood of forest fires to occur in Algeria. Project based on a summative evaluation from UBC's DSCI 100 course

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •