To predict the extent of earthquake damage, we have performed 3 levels of analysis.
1st section: Exploratory Data Analysis (EDA) - uncover hidden trends and insights about data
2nd section: In this section, we aim to evaluate different popular machine learning models and their accuracy in this particular use case of classifying earthquake damage. To do so, we will explore the use of 5 popular machine learning models, tune their hyperparameters to best fit our use case of multi-class classification and identify which model has the best accuracy.
3d section: Evaluate different model optimization methods and their accuracy in this particular use case of classifying earthquake damage. To do so, we will explore the 5 methods and its effect on the accuracy of the Random Forest model. Afterwhich, we will use the most effective methods to build an optimized model to classify earthquake damage.
322th position out of 2392 competitors with a F1 Score of 0.7276 (Top 13% of competitors)