Context

During my exchange semester in Paris, I had to do a machine learning project as part of a course. I chose to build a machine learning model that would predict the occurrence of El Niño events using global sea surface temperatures (SST), as we know that El Niño events occur when there is a temperature increase of at least 0.5 degrees Celsius over five overlapping three-month periods. 

Approach

The data was filtered to retain only the mean temperatures in the Niño 3.4 region. After trying different models (linear regression, random forest, neural networks, ...) to predict the occurrence of El Niño, it turned out that the best results were obtained by a Support Vector Machine (SVM) model combined with Linear Discriminant Analysis (LDA).