Final Project
Final Project
1. EDA.
2. Build 2 simple models with all features.
3. Select important features based on your correlation plot (or any method) and then build 2
or 3 models (Use Optuna or GridsearchCV).
4. Report your final model with the best metric of your choice.
1. EDA
2. Split: X (features) and y (target).
Part 1:
2.1. Work with X (ignore y): perform k-means clustering using any method to select the best k.
2.2. Perform PCA on X by selecting 2 principal components and then report how much variance
can be explained by the model.
Part 2:
2.3. Use both X and y to build at least 3 models and then report the best model (Plot ROC curve).
1. Process column ‘Amount’. Convert to the correct decimal format by first removing ‘.’ and
then replacing ‘,’ with ‘.’ Finally, convert string to float.
2. Process column ‘Date’
3. Build a simple Prophet model and then tune some hyperparameters.