The project aims at predicting the occurence of fire in a future date in the region of Algeria based on the previous data. Based on the weather data of two regions in Algeria, the project aims at training the model and testing that model for classification to predict the occurence of fires in future.
In this project, we had the opportunity to explore a wide range of supervised learning algorithms going from simple algorithms such as Linear Perceptron and Nearest Means Classifier to more complex models such as Support Vector Machines and Random Forests and we got to see how they work the given data. We got to see how simple operations such normalization, go a long way in improving the performance of even the most basic models. We also got to perform a few different feature engineering techniques and they help improve the model performance. After running through these numerous algorithms, we see that the best performance is achieved with combination of PCA and SVM.