On this 3-part collection I’ll cowl most essential questions which are requested in information science interviews and options. I might counsel readers to first reply the questions by themselves and solely then undergo options offered. Half 1 focusses extra on theoretical foundations.
- What’s Batch Normalization and its makes use of in Deep studying?
- Clarify why L1 regularization can result in sparsity however L2 regularization can’t?
- How would you deal with an Imbalanced information if oversampling/underneath sampling isn’t an choice?
- Why is log transformation generally utilized to focus on variables in regression fashions?
- Clarify distinction between Precision, Recall and Accuracy?
- Clarify Bias-Variance tradeoff in Machine Studying. How does it impression mannequin efficiency?
- What’s the Curse of Dimensionality, and the way can or not it’s mitigated in high-dimensional datasets?
- Clarify the idea of p-values in speculation testing. Why is a decrease p-value thought-about vital?
- Clarify the idea of Confidence Intervals (CI). How do you interpret a 95% CI for a inhabitants imply?
- Clarify the idea of the ROC Curve. What does the AUC worth characterize?
