Decision Trees Quant Interview Guide
Decision trees quant interview guide for split criteria, depth, pruning, categorical features, instability, examples, and prompts.
Candidates explaining tree splits, overfitting, and interpretability.
Trees split the feature space
A decision tree repeatedly chooses feature thresholds that improve a criterion such as impurity or loss. The result is easy to inspect but easy to overfit.
Depth controls complexity
Deeper trees can capture interactions but may memorize noise. Interview answers should mention max depth, minimum samples, pruning, and validation.
Concrete example
A tree might split trades by volatility, spread, and time of day. The splits are interpretable, but they need stability checks across periods.
Categorical and missing values need care
Encoding choices, missing-value handling, and rare categories can change splits. State assumptions rather than treating the implementation as magic.
Common mistakes
Candidates often praise interpretability without discussing instability. A single tree can change substantially when the sample changes.
Practice the pattern
Use the LeetQuidity curriculum and calibration to turn this topic into a focused practice plan.