Building and Optimizing a Classification Model for Trading

Load the data, preprocess and clean it

All data is loaded and the charts are displayed correctly.

There are no NaNs in the final DataFrame.

Engineer new features

All the additional features are created correctly.

Detect and eliminate multicollinearity in the features

The correlation matrix heatmap displays corectly.

The right features were dropped.

Define and interpret a reasonable baseline model/score

The baseline score has been calculated correctly.

Train and cross-validate an AI model

The data was split correctly.

5-fold cross-validation was performed correctly and the learning curves display correctly.

Interpret learning curves and detect over-/under-fitting

The learning curves for "manual" tuning of the max_depth hyperparameter are correct.

The student's answers regarding the interpretation of learning curves are correct.

Tune model hyperparameters with GridSearch

GridSearch was performed correctly and the best hyperparameters were chosen.

The student's answers regarding the results of GridSearch are correct.

Evaluate and interpret AI model training results

Performance evaluation metrics are calculated and interpreted correctly.

The student's answer regarding the comparison of the model's performance to the baseline model are correct.

Evaluate and interpret feature importance

| The right features are dropped and the new model is trained correctly.
The results are interpreted correctly. |

Detect and address over-/under-fitting

The student's answers regarding next steps to pursue are reasonable.