Forest fires are a very important issue that negatively affects climate change. Typically, the causes of forest fires are those oversights, accidents and negligence committed by individuals, intentional acts and natural causes. The latter is the root cause for only a minority of the fires.

Their harmful impacts and effects on ecosystems can be major ones. Among them, we can mention the disappearance of native species, the increase in levels of carbon dioxide in the atmosphere, earth’s nutrients destroyed by the ashes, and the massive loss of wildlife. 

Data mining techniques can help in the prediction of the cause of the fire and, thus, better support the decision of taking preventive measures in order to avoid tragedy. In effect, this can play a major role in resource allocation, mitigation and recovery efforts. 

The ICFN - Nature and Forest Conservation Institute has the record of the list of forest fires occurred in Portugal for several years.  For each fire, there is information such as the site, the alert date/hour, the extinction date/hour, the affected area and the cause type (intentional, natural, negligent, rekindling or unknown).

In the files fires2015_train.csv and fires2015_test.csv, you have data on reported forest fires during 2015, for which the cause is known.

**The goal of this practical assignment is to build a machine learning model to predict the cause type of a forest fire**.

Further details on this practical assignment can be found [here](https://www.dcc.fc.up.pt/~rpribeiro/aulas/DMI2021/material/DMI-PracticalAssign_2021.pdf).