Combining Signals for Enhanced Alpha
Features and Labels
| Criteria | Meet Specification |
|---|---|
|
What do you observe in the rolling autocorrelation of labels shifted? |
Describe the relationship between the shifted labels. |
|
Train/Valid/Test Splits |
Correctly implement the
|
Random Forests
| Criteria | Meet Specification |
|---|---|
|
Why does dispersion_20d have the highest feature importance, when the first split is on the Momentum_1YR feature? |
Describe why dispersion_20d has the highest feature importance, when the first split is on the Momentum_1YR feature. |
|
1) What do you observe with the accuracy vs tree size graph? 2) Does the graph indicate the model is overfitting or underfitting? Describe how it indicates this. |
Describe how the accuracy changes over time and what indicates the model is overfitting or underfitting. |
Overlapping Samples
| Criteria | Meet Specification |
|---|---|
|
Drop Overlapping Samples |
Correctly implement the
|
|
Use BaggingClassifier's max_samples |
Correctly implement the
|
|
Build an ensemble of non-overlapping trees: OOB Score |
Correctly implement the
|
|
Build an ensemble of non-overlapping trees: Non Overlapping Estimators |
Correctly implement the
|